In the Name of God Shiraz Journal of System Management (SJSM

Transcription

In the Name of God
Shiraz Journal of System Management (SJSM)
Director-In-Charge: Seyed-Javad Iranban
Editor-In-Chief: Seyed-Mohammad Seyed-Hosseini
Editorial Board:










Afshar Kazemi, Mohammad Ali (Ph.D.) – Shiraz Branch Islamic Azad University,
Shiraz, Iran.
Aminbeidokhti, Ali Akbar (Ph.D.) – Semnan University, Semnan, Iran.
Iranban, Seyed Javad (Ph.D.) – Shiraz Branch Islamic Azad University, Shiraz, Iran.
Kazazi, Aboulfazl (Ph.D.) – Allameh Tabataba'i University, Tehran, Iran.
Khatami Firoozabadi, Seyed Mohammad Ali (Ph.D.) – Allameh Tabataba'i University,
Tehran, Iran.
Maroofi, Fakhraddin (Ph.D.) – University of Kurdistan, Sanandaj, Iran.
Radfar, Reza (Ph.D.) – Shiraz Branch Islamic Azad University, Shiraz, Iran.
Sanoubar, Nasser (Ph.D.) – University of Tabriz, Tabriz, Iran.
Seyed-Hosseini, Seyed-Mohammad, (Ph.D.) – Iran University of Science and
Technology, Tehran, Iran.
Toloie Eshlaghy, Abbas, (Ph.D.) – Shiraz Branch Islamic Azad University, Shiraz, Iran.
Address: Shiraz Journal of System Management Office, Islamic Azad University,
Pardis Sadra, Shiraz, Iran. www.sjsmjournal.com
[email protected]
ISSN: 2322-2301
Shiraz Journal of System Management
Vol. 2, No. 1, Ser. 5, (2014), 1-19
Using Genetic Algorithm to Robust
Multi Objective Optimization of Maintenance
Scheduling Considering Engineering Insurance
Somayeh Molaei∗
Department of Industrial Engineering,
Amirkabir University of Technology,
Iran, Tehran
[email protected]
Mir Mahdi Seyed Esfahani
Iran, Tehran
[email protected]
Akbar Esfahanipour
Iran, Tehran
[email protected]
Abstract. Efficient and on-time maintenance plays a crucial role in
reducing cost and increasing the market share of an industrial unit. Preventive maintenance is a broad term that encompasses a set of activities
aimed at improving the overall reliability and availability of a system
before machinery breakdown. The previous studies have addressed the
scheduling of preventive maintenance. These studies have computed the
time and the type of preventive maintenance by modeling the total cost
related to it. Todays the engineering insurance is an appropriate and
durable protection for reducing the risks related to the industrial machinery. This kind of insurance covers a part of maintenance costs. Previous researches did not consider the effect of engineering insurance
on maintenance scheduling while it affects the total cost function of
∗
Received: January 2014; Final Revision: April 2014
Corresponding author
1
2
S. Molaei, M. M. Seyed Esfahani, and A. Esfahanipour
maintenance scheduling seriously. Given the above-mentioned remarks,
this paper introduces for the first time a new scheduling of preventive
maintenance with considering total cost and total reliability of the system in which the effect of engineering insurance has been taken into
account. Due to the uncertainty in the input parameters, which are
very common in application, the paper proposed the application of robust design approaches. To solve this multi objective model, first it has
been transformed into a single objective model by using global criterion and the resultant model is solved through genetic algorithm. The
results show the magnitude effect of engineering insurance on maintenance scheduling. Therefore, neglecting the importance of engineering
insurance leads to an inefficient scheduling maintenance.
Keywords: Preventive maintenance; engineering insurance; robust optimization; global criterion method; genetic algorithm; Pareto set solutions.
1.
Introduction
In the current competitive environment managers of manufacturing and
service organizations try to make their organizations competitive by providing timely delivery of high quality products. For this purpose, maintenance plays a key role in reducing cost, minimizing equipment downtime,
improving quality, increasing productivity, providing reliable equipment,
and as a result achieving organizational goals and aims [1]. Despite playing such an important role, the maintenance costs are a major part of the
total operating costs of all manufacturing or production plants. Depending on the specific industry, these costs may represent between 15 and
60 percent of the costs of goods produced [2]. Recent surveys of maintenance management effectiveness indicate that one-third of all maintenance costs is wasted because of unnecessary or improperly carried out
maintenance. The result of ineffective maintenance management represents a loss of more than 60$ billion each year [3].
Therefore, using an efficient maintenance strategy can reduces the total costs of production in manufacturing or production plants. Industrial
and process plants typically use two types of maintenance management,
run-to-failure or preventive maintenance [4]. Run-to-failure or corrective maintenance is a reactive management technique that waits for ma-
Using Genetic Algorithm to Robust Multi ...
3
chine or equipment failure before any maintenance action is taken. It
is also the most expensive method of maintenance management. While
the Preventive maintenance encompasses a set of activities aimed at improving the overall reliability and availability of a system before fault
occurs. Preventive maintenance is further divided into periodic maintenance and predictive maintenance. Periodic maintenance is a time-based
maintenance consists of periodically inspecting, servicing and cleaning
equipment and replacing parts to prevent sudden failure and process
problems. Predictive maintenance is a method in which the service life
of important part is predicted based on inspection or diagnosis, in order to use the parts to the limit of their service life. Compared with
periodic maintenance, predictive maintenance is condition-based maintenance. In general, preventive maintenance activities are categorized in
one of two ways, component maintenance or component replacement. It
is clear that preventive maintenance involves a basic trade-off between
the costs of conducting maintenance and replacement activities and the
cost savings achieved by reducing the overall rate of occurrence of system failures. Designers of preventive maintenance schedules must prioritize these individual costs to reduce the overall cost of system operation. They may also be interested in maximizing the system reliability,
subject to some sort of budget constraint. In periodic maintenance the
optimum scheduling is very necessary.
Many researches have been optimized maintenance scheduling economically. Reference [5] determined optimal cost of maintenance policies
by defining the average cost rate of system operation, in this study it
is assumed that an increasing failure rate is based on the Weibull distribution function. Reference [6] developed a model to minimize the
total costs related to preventive maintenance schedules. Exact algorithms reach exact optimal solutions of mathematical models, while
approximation algorithms seek an approximation that is close to the
true optimal solutions. Reference [7] presented a model that optimizes
the preventive maintenance scheduling in semiconductor manufacturing operations. They optimized this model via a mixed-integer linear
programming model. Reference [8] presented a preventive maintenance
optimization model in order to minimize the total maintenance costs
4
in a production system. Reference [9] determined an optimal preventive maintenance schedule by considering the time value of money in all
future costs. Reference [10] defined the summation of maintenance activities cost along with cost of unsupplied demand due to failures of components in the objective function to optimize maintenance strategy. Reference [11] presented a model in order to optimize the maintenance policy
for a component with random failure rate. Reference [12] presented an
optimization model to schedule a preventive maintenance. He considered
the total cost relating to operations as the objective function and solved
the model using Bender’s decomposition. Reference [13] presented two
mixed-integer linear programming models for preventive maintenance
scheduling problems and use CPLEX to implement the optimization
models for a case study of railway maintenance scheduling. Reference
[14] developed an age based nonlinear optimization model to determine
the optimal preventive maintenance schedule for a single component system. Reference [15] developed three nonlinear optimization models: one
that minimizes total cost subject to satisfying a required reliability, one
that maximizes reliability at a given budget, and one that minimizes
the expected total cost including expected breakdown outages cost and
maintenance cost.
Because of complexity of maintenance scheduling, metaheuristic algorithms have been used in several researches [16]. Reference [17] used
genetic algorithms with simulated annealing in order to optimize a largescale and long-term preventive maintenance and replacement scheduling
problem. Reference [18] used an ant colony algorithm to optimize the
maintenance scheduling. Reference [19] proposed several techniques for
representing the decision variables in preventive maintenance scheduling models and used heuristics and metaheuristics optimization algorithms. Reference [20] developed a novel multi-objective genetic algorithm to optimize preventive maintenance schedule problems. Reference
[21] presented a production planning model considering preventive maintenance which minimize the completion time of jobs and downtime of
machines. Reference [22] solved the previous model using genetic algorithm and simulated annealing algorithm. Reference [23] presented
comprehensive research in area of integrating preventive maintenance
5
and production scheduling and computed the Pareto set solution.
Insurance is a financial topic of paramount importance for every industry and is designed to protect the financial well-being in the case
of unexpected loss. One of the new types of general insurance products is engineering insurance. This type of insurance is an appropriate
and durable protection for reducing the risks related to the industrial
machinery. It covers a part of maintenance costs.
Previous researches did not consider the effect of engineering insurance on maintenance scheduling optimization, while it affects the total
cost function of maintenance scheduling seriously. So it is necessary to
rewrite the cost function considering the cost of maintenance which is
compensated by engineering insurance. This paper introduces for the
first time a new scheduling of preventive maintenance considering total
cost and total reliability of system in which the effect of engineering
insurance has been taken into account. Due to the uncertainty in the
input parameters, which are very common in application, the paper
proposed the application of robust design approaches. Because of the
complexity of the proposed model genetic algorithm is used to compute
the Pareto set solution. The organization of this paper is as follows. In
Section II, machinery breakdown insurance is illustrated; in Section III
the preventive maintenance scheduling model is presented. Section IV
demonstrates the structure of the robust design. Section V illustrates
the multi objective optimization problem. In Section VI the proposed
genetic algorithm which is used to solve the optimization problem is explained in details and finally, Section VII concludes the research with
future researches
2.
Machinery Breckdown Insurance
Engineering insurance refers to the insurance that provides economic
safeguard to the risks faced by the ongoing construction project, installation project, and machines and equipment in project operation. Depending on the project, it can be divided into construction project all risks
insurance and installation project all risks insurance; depending on the
attribute of the object, it can be divided into project all risks insurance,
6
and machinery breakdown insurance. Machinery Breakdown Insurance
is designed to provide cover against unforeseen and sudden physical loss
or damage to the machinery by any cause subject to excepted risks. Machines are an integral part of all manufacturing and industrial units
engaged in production of industrial or household goods. These may be
large industrial establishments or small and medium enterprises and any
unexpected accident or breakdown to their critical machinery brings it
to a standstill adversely affecting business and causing a financial strain
towards repair or replacement of the affected machinery. The Machinery Breakdown Insurance offers coverage to organization against these
sudden unforeseen accidents or events.
3.
Preventive Maintenance Scheduling
Variables
N: Number of components
T: Number of periods
J: Number of intervals
λi : Scale parameter of component i
βi : Shape parameter of component i
Z: Fixed cost of system
Tri : Time required to replace component i
TM i : Time required to maintain component i
fi (t): Probability distribution function
fij : Total cost due to failure of component i in period j
P i: Premium of component i
Reli,t : Reliability of component i at the start of period j
δi : Percent of premium which is paid when the component i is replaced
τi : Percent of maintenance cost which is compensated by engineering
insurance when the component i is maintained.
Fi : Unexpected failure cost of component i
Mi : Maintenance cost of component i
Ri : Replacement cost of component i
xij : Effective age of component i at the start of period j
yij : Effective age of component i at the end of period j
16
tal cost of system is modeled considering engineering insurance. After
that, a numerical example is solved in which the cost parameters are
considered not deterministic, therefore; a robust model of maintenance
is proposed. To compute the optimal results, first the multi objective is
transformed to single objective using global criterion method; then the
genetic algorithm is used and Pareto optimal solution are computed.
Finally the effect of engineering insurance is investigated on the optimal preventive maintenance and replacement scheduling. Results show
that engineering insurance affects the optimal maintenance scheduling
seriously, therefore ignorance of engineering insurance in maintenance
scheduling problem lead to inefficient scheduling plan. In future researches the effect of preventive maintenance on premium of insurance
can be investigated.
References
[1] Bashiri, M., Badri, H., and Hejazi, T. H. (2011), Selecting optimum maintenance strategy by fuzzy interactive linear assignment method. Applied
Mathematical Modeling, 35 (1), 152-164.
[2] Bevilacqua, M. and Braglia, M. (2000), The analytic hierarchy process applied to maintenance strategy selection. Reliability Engineering & System
Safety, 70 (1), 71-83.
[3] Mobley, R. K. (2002), An introduction to predictive maintenance.
Butterworth-Heinemann.
[4] Li, J. R., Khoo, L. P., and Tor, S. B. (2006), Generation of possible multiple components disassembly sequence for maintenance using a disassembly
constraint graph. International Journal of Production Economics, 102 (1)
51-65.
[5] Canfield, R. V. (1986), Cost optimization of periodic preventive maintenance. Reliability, IEEE Transactions, 35 (1), 78-81.
[6] Panagiotidou, S. and Tagaras, G. (2007), Optimal preventive maintenance
for equipment with two quality states and general failure time distributions. European journal of operational research, 180 (1), 329-353.
17
[7] Yao, X., Fu, M., Marcus, S. I., and Fernandez-Gaucherand, E. (2001), Optimization of preventive maintenance scheduling for semiconductor manufacturing systems: models and implementation. In Control Applications,
Proceedings of the 2001 IEEE International Conference, 407-411.
[8] Charles, A. S., Floru, I. R., Azzaro-Pantel, C., Pibouleau, L., and
Domenech, S. (2003), Optimization of preventive maintenance strategies
in a multipurpose batch plant: application to semiconductor manufacturing. Computers & chemical engineering, 27 (4), 449-467.
[9] Usher, J. S., Kamal, A. H., and Syed, W. H. (1998), Cost optimal preventive maintenance. IIE transactions, 309 (12), 1121-1128.
[10] Levitin, G. and Lisnianski, A. (2000), Short communication optimal replacement scheduling in multi?state series-parallel systems. Quality and
Reliability Engineering International, 16 (2), 157-162.
[11] Jayakumar, A. and Asgarpoor, S. (2004), Maintenance optimization of
equipment by linear programming. In Probabilistic Methods Applied to
Power Systems, 145-149.
[12] Canto, S. P. (2008), Application of Benders’ decomposition to power plant
preventive maintenance scheduling. European journal of operational research, 184 (2), 759-777.
[13] Budai, G., Huisman, D., and Dekker, R. (2005), Scheduling preventive
railway maintenance activities. Journal of the Operational Research Society, 57 (9), 1035-1044.
[14] Shirmohammadi, A. H., Zhang, Z. G., and Love, E. (2007), A computational model for determining the optimal preventive maintenance policy
with random breakdowns and imperfect repairs. Reliability, IEEE Transactions, 56 (2), 332-339.
[15] Tam, A. S. B., Chan, W. M., and Price, J. W. H. (2006), Optimal maintenance intervals for a multi-component system. Production Planning and
Control, 17 (8), 769-779.
[16] Moghaddam, K. S. and Usher, J. S. (2011), Preventive maintenance and
replacement scheduling for repairable and maintainable systems using dynamic programming. Computers & Industrial Engineering, 60 (4), 654665.
18
[17] Kim, H., Nara, K., and Gen, M. (1994), A method for maintenance
scheduling using GA combined with SA. Computers & Industrial Engineering, 27 (1), 477-480.
[18] Samrout, M., Yalaoui, F., Chtelet, E., and Chebbo, N. (2005), New methods to minimize the preventive maintenance cost of series-parallel systems
using ant colony optimization. Reliability engineering & system safety, 89
(3), 346-354.
[19] Limbourg, P. and Kochs, H. D. (2006), Preventive maintenance scheduling
by variable dimension evolutionary algorithms. International journal of
pressure vessels and piping, 83 (4), 262-269.
[20] Quan, G., Greenwood, G. W., Liu, D., and Hu, S. (2007), Searching for
multiobjective preventive maintenance schedules: Combining preferences
with evolutionary algorithms. European Journal of Operational Research,
177 (3), 1969-1984.
[21] Moradi, E., Fatemi Ghomi, S. M. T., and Zandieh, M. (2011), Bi-objective
optimization research on integrated fixed time interval preventive maintenance and production for scheduling flexible job-shop problem. Expert
systems with applications, 38 (6), 7169-7178.
[22] Naderi, B., Zandieh, M., and Aminnayeri, M. (2011), Incorporating periodic preventive maintenance into flexible flow shop scheduling problems.
Applied Soft Computing, 11 (2), 2094-2101.
[23] Berrichi, A., Yalaoui, F., Amodeo, L., and Mezghiche, M. (2010), BiObjective Ant Colony Optimization approach to optimize production and
maintenance scheduling. Computers & Operations Research, 37 (9), 15841596.
[24] Smith. C. O. (1976), Introduction to Reliability in Design, first edition,
McGraw-Hill, New York.
[25] Yun, W. Y. and Kim, J. W. (2004), Multi-level redundancy optimization
in series systems. Computers & Industrial Engineering, 46 (2), 337-346.
[26] Kouvelis, P. and Yu, G. (1997), Robust discrete optimization and its applications, 14, Springer.
[27] Serafini, P. (1994), Simulated annealing for multi objective optimization
problems. In Multiple criteria decision making, 283-292).
19
[28] Suppapitnarm, A., Seffen, K. A., Parks, G. T., and Clarkson, P. J. (2000),
A simulated annealing algorithm for multiobjective optimization. Engineering Optimization, 33 (1), 59-85.
Vol. 2, No. 1, Ser. 5, (2014), 21-37
A Study on Effective Factors on New Product
Development with an Emphasis on Fuzzy
Hierarchical Analysis Approach
Javad Iranban∗
Department of Management,
Shiraz Branch, Islamic Azad University,
Shiraz, Iran
[email protected]
Sanaz Davodzadeh
Shiraz Branch, Islamic Azad University,
Shiraz, Iran
Abstract. Nowadays the new product and its importance are considered as an essential strategy for staying in business. Though the hi-tech
industry has focused on value innovation and improving the quality
of the new product development (NPD) process to drive new product performance, new product success has not changed dramatically
over the years. This study presents a novel approach based on structural equation modeling (SEM) and fuzzy analytical hierarchy process
model (FAHP) to explore how value innovation and quality of new product process affect NPD performance.
The survey contained industrial companies which are located in Fars
province. The sample contains 98 people who were selected by random
cluster sampling. The research was done using descriptive and applied
method. Totally, 16 indicators have been collected and were classified
in 3 groups of value innovation, quality and NPD performance. Results
demonstrate that value innovation directly affects NPD performance
and the quality of NPD processes has a nonlinear effect on NPD performance.
Keywords: Value innovation; new product development; fuzzy analytical hierarchy process; Structural Equation Modeling.
∗
Received: February (2014); Final Revision: May (2014)
21
22
1.
J. Iranban and S. Davodzadeh
Introduction
Rapidly changing technologies, intense competition, and dynamic customer needs and wants are rendering existing products obsolete, resulting in shorter product life cycles. The dilemma faced by management
in high technology firms is that while the need for successful NPD is
stronger than ever, new products success has not changed dramatically
over the years [12]. Nowadays organizations have found that mere relying and trusting on traditional competitive factors are not sufficient and
some concepts like speed and flexibility in competition are important
instead. Changes in the tendency towards providing new products and
services to market are the reason of this attitude. On the other hand
should have been an appropriate response to the changing competitive
market. Therefore, the new products will increase the profits of the company. Studies have suggested a wide range of factors that drive new
product performance [5]. Most managers will agree that value innovation enables continuous growth and profits, and plays a vital role in new
product development (NPD) in hi-tech industry [7]. In recent decades,
the technology management literature has emphasized the importance of
value innovation in creating and sustaining competitive advantage and
in rejuvenating the enterprise [1] [2] [6] [7]. Further, the quality of NPD
processes is also a major factor in new product performance [4] [5].
Increasing speed of technology development, short life cycles and
increasing competition and turbulent economic climate of the twentyfirst century increase the importance of product development. The main
source of success in achieving competitive for the company’s in the future is the successful development of new products and continuous improvement. With the increasing in the variety of products and volume of
orders the survival curves of the products will be decreased. For example, a group of products in 1920s have a mean time to mass production
about 28 years old and this period has decreased to 10 years in 1960. Estimates show that the number of new products that will be released in
the next 5 years, will be twice as the products which have been supplied
in last 5 years. This growth led to a 40 percent will lead to increase in
sales and a 30 percent increase in corporate profits. So, changing rules
A Study on Effective Factors on New Product ...
23
of competition in the business have increased the importance of delivering new products to market [3]. According to the study which was
conducted in 1981, about 700 American companies, the results showed
that about a third of the profits of these organizations are due to their
new products while the profits of1970was a third. Due to these factors,
new product development management process also requires the use of
new approaches to management [17].
One of the critics to the development of new products is the complexity and uncertainty of the process of the development of new products. Many researches in this area studies have been considered different
aspects of product value innovation but this knowledge has not been
used in the real-world. Due to the importance of NPD process, this research will try to study the introduction of new product development
and analyze value innovation and quality of NPD process to define their
effect on the NPD performance.
2.
Literature Revien
Development of a new product is a process within that the new product or service is launched to the market and consists of two parallel
paths. The first way involves innovative ideas, product design and product engineering, and the second way involves research and market analysis. The success of this process can be measured through the determination of the indexes which reflects the degree of success or failure of
these indexes. Some of the most important indicators that can help to
the formulation of statements include value innovation, quality of NPD
process, sales, and volume of customers, annual revenue growth and
growing of the products portfolio. This term NPD is applied about the
new products in the world and also about the application of the minimal
change and improvement in existing product [3].
Lifetime of the products in the market is declining and the rate of
product development is expected to double every five years. As a result,
new products that meet customers’ needs and desires is a key factor
in maintaining and improving competitive advantage. Global economy
with market segmentation has significant impact on the development of
24
new products and mass production has less importance than the level
of mass customization.
Every company given the circumstances of its internal and external
environment should perform a new plan for the development of new
products. About the development of new products, there are two types
of internal and external growth. The most important factors affecting
domestic growth includes: value innovation in products, having focus
on the labor market and using for skilled and qualified people in key
posts. Key factors in the external growth includes having appropriate
strategy staffing and financial control [18].
With considering above mentioned factors, the ability to the production and the development of products from external sources is acquired. One of the biggest influences on the approach of Western companies on the product development is derived from the developed concept by NASA in the 1960s which was considering to the management
of complex large-scale defense projects. This approach consists of four
consecutive phases: initial analysis (first phase), definition (Phase II),
design (Phase III), Operations (Phase IV).
There are different studies based on the location and condition of
the product unit in connection with an extensive study of the factors
affecting the success of new product development. Market orientation
and subsequent firm innovation are widely recognized to be essential for
the survival and growth of organizations[1][2]. Value innovations require
an organizational commitment to create a strong momentum for insidein changes and advances in bringing inside-out value creating outcomes
and attaining superior positions in the competition race [14]. Continuing
success in delighting the customer, in turn, drives sustained increase in
enterprise value. Aiman-Smith et al. (2005) defined value innovation as
that innovation which occurs when organizational members are working on identifying better ways to serve their current customers, and on
identifying new markets [1]. Dikmen et al. (2005) pointed out that value
innovation is not a competition-based view of the firm, but instead is
an endogenous growth theory and a resource-based view, where growth
and innovation come from within the organization itself. Thus, innovations emerge from knowledge accumulated within the organization and
25
resource recombination chosen by the firm to produce a service/product
[6].
Many organizations believing that technology innovation alone can
create new wealth [7]. Irwin, Hoffman, and Lamont (1998) used a resourcebased view to demonstrate the positive relationship between technological innovations and organizational performance [11]. Hurley and Hult
(1998) showed that positive relationships between organizational innovations influenced the potential for good performance [10]. Aragn-Correa et
al. (2007) showed the positive effect of firm innovation on performance
[2]. In view of the positive relationships seen in previous research, we
thus hypothesize:
Hypothesis 1. Higher value innovation will have a positive effect on
NPD performance.
Cooper (1996) argued that a high quality new product process to
guide product innovations from idea to launch is a critical success factor. New product processes have been found to fail for a number of
reasons. First, inadequate up-front homework has been found to be a
major cause of failure in product development [5]. Second, failure to
define the product before development begins can cause both new product failure and serious delays in the development cycle. Third, many
failed projects were moved too far into development without serious
scrutiny. The lack of tough Go/Kill decision points meant too many
product failures, resources wasted on the wrong projects, and a lack of
focus. Fourth, an emphasis on quality-of-execution in many firms came
about after internal studies revealed that too many projects suffered
from weak, inconsistent work, with some of the most deficient areas
being the market-related processes and routines at improving quality
of execution of key tasks and activities throughout the process. Fifth,
many companies discovered that not only was the quality of work unacceptable, but needed work such as market analysis, business assessment,
and customer research, were simply not done or displayed hasty corner
cutting. Finally, the new product process was inflexible or overly formalized, with stages and decision points that could be not skipped or
combined, becoming a straightjacket for the [5].
26
Cooper and Kleinschmidt (2007) demonstrated that the strongest
driver of profitability is the existence of a high-quality and rigorous new
product process that emphasizes up-front homework, tough Go/Kill decision points, sharp early product definition, and flexibility. By contrast,
merely having a formal new product process has no impact at all on performance. Cooper (1996) demonstrated that a high quality new product
process had the strongest impact on new product performance. Hence,
since high quality new product processes appear to yield positive NPD
performance, we construct the following hypothesis:
Hypothesis 2. A high quality of NPD process will have a positive
impact on NPD performance.
3.
Conceptual Model of Research
After analyzing value innovation and the quality of NPD process, model
in figure 1 chased as a conceptual model and the base of this research. In
this model we will reviewed the hypotheses, considering three main
variables of value innovation, quality of NPD process and NPD performance. The research hypotheses explore how value innovation and
quality of new product process affect NPD performance. To measure
value innovation 9 indicators including: Customer orientation, Agile
decision-making, Business intelligence, Open communication, Empowerment, Business planning, Organization learning Meaningful work, Risktaking culture. Quality of NPD is measured by four indicators: Solid
up-front homework, Sharp, early product definition, Quality of execution throughout, A flexible process. And for NPD performance variable
there are three indicators including Customer performance, Technology
performance, and Market performance.
4.
Research Methodology
Since the purpose of this study is the analysis of the effects of “value
innovation” and “product quality” on new product performance this
research is applicable and descriptive. This study consists of all indus-
35
beyond the current demand and to attract mass new customers who did
not exist before.
According to the results of fuzzy hierarchical analysis, customer orientation is the variable that has most significant impact in value innovation
variables, Therefore it is recommended to organizations to focus on develop NPD based on customer knowledge management framework, the
customer’s knowledge should be elicited and converted to a pattern of
consumer’s need towards the products attributes. And about the quality variable, Quality of execution throughout has the highest weight so
recommended to organizations, investment in quality control segment.
7.
Suggestions for Furder Research
According to the results, researches can be done in this regard:
1. Other factors affect new product development performance.
2. Predicting the performance of new product development using structural equation modeling.
3. Predicting value innovation performance with neural networks and
their comparison.
4. There should be a framework for measuring value innovation in firm.
5. Provide indicators to measure value innovation activities.
References
[1] Aiman-Smith, L., Goodrich, N., Roberts, D., and Scinta, J. (2005), Assessing your organization’s potential for value innovation .Research Technology Management, 48 (2), 225-246.
[2] Aragon-Correa, J. A. and Garcia-Morales, V. J., Cordon-Pozo ,E.(2007),
Leadership and organizational learning’s role on innovation and performance: Lessons from Spain. Industrial Marketing Management, 36 (4),
349-359.
[3] Barclay, K. 2006, Between modernity and primitively: Okinawan identity
in relation to Japan and the south pacific. Nations and Nationalism, Vol.
12, No. 1, 117-138.
36
[4] Cooper, R. G. (1996), Overhauling the new product process. Industrial
Marketing Management, 25 (6), 465-482.
[5] Cooper, R. G. and Kleinschmidit, E. J. (2007), Winning businesses in
product development: The critical success factors. Research Technology
Management, 50 (3), 52-66.
[6] Dikmen, L., Birgonul, M. T., and Artuk, S. U. (2005), Integrated framework to investigate value innovations. Journal of Management in Engineering, 21 (2), 81-90.
[7] Dillon, T. A., Lee, R. K ., and Matheson, D. (2005), Value innovation
:Passport to wealth creation. Research Technology Management, 48 (2),
22-36.
[8] Garcia-Morales, V. J., Ruiz-Moreno, A., and Loren’s-Montes, F. J.
(2007), Effects of Technology absorptive capacity and technology proactivity on organizational learning, innovation and performance: An empirical examination. Technology Analysis and Strategic Management, 19 (4),
527-558.
[9] Haider A. (1388), Structural Equation Modeling using lisrel software.
Third edition. SAMT.
[10] Hurley, R. F. and Hult, G. T. (1998), Innovation, market orientation,
and organizational learning: An integration and empirical examination.
Journal of Marketing, 62 (3), 42-54.
[11] Irwin, J. G., Hoffman, J. J., and Lamont, B. T. (1998), The effect of the
acquisition of Technological innovations on organizational performance:
A resource-based View. Journal of Engineering and Technology Management, 15 (1), 25-54.
[12] Joshi, A. W. and Sharma, S. (2004), Customer knowledge development:
Antecedents and impact on new product performance. Journal of Marketing, 68 (4), 47-59.
[13] Kline, R. B. (2005), Principles and practice of structural equation modeling (2nd Ed.). New York: The Guilford Publications. Inc.
[14] Mohanty, R. P. (1999), Value innovation perspective in Indian organizations. Participation and Empowerment: An International Journal, 7 (4),
88-103.
37
[15] Sethi, R. and Iqbal, Z. (2008), Stage-Gate controls, learning failure, and
adverse effect on novel new products. Journal of Marketing, 72 (1), 118134.
[16] Taghavi fard, T. and Akhbari, M. (1386), New product development process. Tadbir scientific-educational publication, 184.
[17] Page, A. and Jones, R. (1989), Business growth How to achieve and sustain. Leadership and Organization Development Journal, 10 (2), 1-55.
[18] Varela, J. and Benito, L. (2005), New product development process in
Spanish firms: Typology,antecedents and technical/marketing activities.
Technovation, 25 (4), 395-405.
Vol. 2, No. 1, Ser. 5, (2014), 39-49
Predicting Customer Churn Using CLV in
Insurance Industry
Vahid Dust Mohammadi∗
Tarbit Modares University,
Tehran, Iran
[email protected]
Amir Albadvi
Tehran, Iran
[email protected]
Babak Teymorpur
Tehran, Iran
[email protected]
Abstract. Today, increased level of customer awareness caused them
to access to the other suppliers easily and they can get their services
from the competitors with similar or even better quality and same price.
Therefore, focusing on customers and preventing them to leave, has been
the most important strategy for any company. Researches have shown
that retaining former customers is cheaper than attracting new ones. In
the proposed model in this article we first identified important factors
causing customers in insurance industry, to have a specific behavior by
using a k-means clustering algorithm, and then we tried to predict the
future behavior of them by a logistic regression. Our model accuracy is
98%.
Keywords: Customer churn, customer lifetime value, k-means clustering, logistic regression, insurance industry.
∗
Received: January (2014); Final Revision: May (2014)
39
40
1.
V. Dust Mohammadi, A. Albadvi, and B. Teymorpur
Introduction
Changing economic and social characteristics caused organizations to do
things in a different way.[1] Unlike the past, today customers are determining success factors of organizations. On the other hand, using new
technologies like internet made making new customers and finding new
suppliers easier and cheaper. In this situation new entrants can enter
into the competing market with very low costs and lots of substitutes
can be found just by visiting some sites in the internet. All of these
features of using the internet have made the customers stronger than
before, customers who can choose between different suppliers and different substitutes. On the other hand survining in the competing market is
more complicated for organizations.[2] Organizations have known that
their most important asset is their customers, so all of their effort is
to reach to a larger market share and creating value for more valuable
customers. One of the issues that organizations encounter is the way
they should communicate with their customers, in the other words; they
are supposed to maintain their old customers in addition to making new
ones. In this situation the more successful organization is the one who
can maintain its valuable customers and by making appropriate policies preventing them to leave ,because maintaining the old customers
is more cheaper than making new ones.[3] So, different researches have
been done recently in the field of detecting organization’s valuable customers are very important and helpful to find good ways of preventing
customers to churn.
In the next section the importance of churn analysis from the viewpoint of different researches is going to be studied, then we explain the
characteristics of data we have used in this article and in the last section
we propose our predicting model for analyzing customer behavior.
2.
Customer Relationship Management (CRM)
Customer relationship management includes processes and systems which
support a long and valuable relationship with special customers. [4] Customer behavior records and information technology is the base of any
Predicting Customer Churn Using CLV ...
41
CRM strategy. The increasing use of the internet and its related technologies has made marketing more easily and the way customers communicating with organizations are being changed due to these services. Although CRM is one of the important business strategies, but there is
not already an international specified definition for it. [4, 5]
In different researches different definitions of CRM can be found:
• An organizational trend for analyzing and affecting customer behavior
along a valuable period of communication for better acquisition and
maintaining of loyal customers. [5]
• Strategic using of information, processes, technologies and people for
managing the relationship with customers with organizations (including
marketing, sales, service and support) during the customer lifecyle. [6]
• A general strategy and process of acquisition, maintaining and communicating with customers for creating value for organization and customer. CRM includes marketing, sales, services and organization’s supply chain for better performance and efficiency in creating value for
customers. [7]
CRM includes customer identification, customer attraction, customer
retention and customer development [8]. Analyzing customer churn is
one of the issues of CRM which is being classified in customer retention
field.
Today in industrial countries the number of service firms is increasing and Iran is not an exception. The growth rate of the number of
insurance companies is impressive in Iran. There are 20 insurance companies (one of them acts as a public company and the rest belongs to
the private sector), 15,200 insurance agents and 270 brokerages act in
Iran insurance industry [9]. The penetration rate of insurance in Iran
is 1.5% of GDP and if social insurance premiums, pension funds and
supportive insurance are counted in this ratio it is 4.6% of GDP [9]. So
the importance of insurance industry cannot be denied in the country.
The existence of different insurance companies and new entrants to the
insurance industry in one hand, and using new technologies like the internet, which made more informed customers and competitors, in the
other hand, made competition for customer attraction more intense in
43
According to the above definitions, CLV can be defined as the collection of revenues from customers of the organization along their interaction period, which attraction, sale and service costs are subtracted from
the, and is declared in terms of time value of money. What is obvious in
the above definitions is the historical attitude to the customer purchases
and they don’t define any potential for the customer, this defect has
been corrected in the new models. [13] CLV is calculated as:
n
(Ri − Ci )
CLV =
(1 + d)i−0.5
(1)
i=1
Ri = the amount of customer revenue that the organization gains;
Ci = the cost of services that the organization provides; D= the amount
of discount rate;
I= the number of periods that customer have transaction with the organization.
Logistic regression is one of the applicable techniques for analyzing
classified data. For example, if the result of an experiment is defined as
loose or win, then the dependent variable is not continues and will be
a categorical variable. One kind of logistic regressions is binary logistic
regression which it has two classes of dependent variables. [14]
This model is defined as:
p
logit(p) = ln( 1−p
) = β0 + β1 x1,i + ... + βk xk,i
p = pr(yi = 1|X) =
eβ0 +β2 x2,i +...+βk xk,i
1+eβ0 +β2 x2,i +...+βk xk,i
(2)
(3)
β0 = the constant of the equation
β= the coefficient of the predictor variables
3.
Research Data Features
In this article the data of four years of third party insurance of one
of the insurance companies of Iran used for generating a predictive
47
As shown in the table 3 attributes were not significantly correlated
and no multicollinearity existed between them (standard error for all
attributes was less than 2). Also, according to p-value column null
hypothesis for logistic regression model was rejected for all attributes
(p < 0.05). These attributes all together had a meaningful share in predicting the probability of churn.
The coefficients show that which attributes increase or decrease the
probability of leaving the company. Attributes with negative coefficient
has negative impact of not being churned and attributes with positive
coefficients has positive impact in being a churn customer. In the other
words number of installments, payment, discount and the number of
contracts has negative impact of being a churn customer and the rest
attributes has positive impact in being a churn one.
For analyzing the goodness of fit we used Cox & Snell and Nagelkerke
which they were 0.72 and 0.97 in the SPSS output. As shown, 72% to
97% of variability of the dependent variable can be predicted by the
independent variables.
5.
Conclusion
As shown in the previous paragraph, 5 attributes (payment, installment,
value, CLV, number of contracts and discount) are the most important
attributes in the churn management modeling in Iran insurance industry. In the other words, attributes like demographic, application of the
insured car doesn’t have a meaningful impact on customer churn. Among
the extracted attributes number of installment and payments and the
amount of discount have an inverse impact in customer churn and the
rest attributes have a direct impact. Among these attributes having good
policies for discount rate, installment and making customers to have
more contracts can be very helpful for success in churn management.
Acknowledgments:
The research reported herein was supported by the insurance research
center affiliated to the central insurance of the Islamic Republic of Iran.
The data used for this research was provided by the Iran insurance
48
V. Dust Mohammadi, A. Albadvi, and B. Teymorpur
company. We gratefully acknowledge all our sponsors. We also thank
Mr. Hazraty for his helpful suggestions and feedback.
References
[1] Berry, M. J. and Linoff, G. S. (2004), Data mining techniques: for marketing, sales, and customer relationship management: John Wiley & Sons.
[2] Turban, E., Sharda, R., Delen, D., and Efraim, T. (2007), Decision support and business intelligence systems: Pearson Education India.
[3] Tsai, C. F. and Lu, Y. H. (2009), “Customer churn prediction by hybrid
neural networks,” Expert Systems with Applications, vol. 36, pp. 1254712553.
[4] Ling, R. and Yen, D. C. (2001), “Customer relationship management: An
analysis framework and implementation strategies,” Journal of Computer
Information Systems, vol. 41, pp. 82-97.
[5] KhakAbi, S., Gholamian, M. R., and Namvar, M. (2010), “Data mining applications in customer churn management,” in Intelligent Systems,
Modelling and Simulation (ISMS), International Conference on, pp. 220225.
[6] Zhang, Y., Liang, R., Li, Y., Zheng, Y., and Berry, M. (2011), “BehaviorBased Telecommunication Churn Prediction with Neural Network Approach,” in Computer Science and Society (ISCCS), International Symposium on, pp. 307-310.
[7] Benoit, D. F. and Van den Poel, D. (2012), “Improving customer retention
in financial services using kinship network information,” Expert Systems
with Applications, vol. 39, pp. 11435-11442.
[8] Kracklauer, A., Mills, H. D. Q., and Seifert, D. (2004), “Customer
management as the origin of collaborative customer relationship management,” in Collaborative Customer Relationship Management, Ed:
Springer, pp. 3-6.
[9] Kajvary, (2013), “important factors in brand equity in insurance industry
in customer perspective,” insurance journal.
[10] Chu, B. H., Tsai, M. S., and Ho, C. S. (2007), “Toward a hybrid data
mining model for customer retention,” Knowledge-Based Systems, vol.
20, pp. 703-718.
49
[11] Kim, H. S. and Yoon, C. H. (2004), “Determinants of subscriber churn
and customer loyalty in the Korean mobile telephony market,” Telecommunications Policy, vol. 28, pp. 751-765.
[12] Bahman, A. (2014), “modeling CLV in insurane industry,” engineering,
tarbiat modares, Iran.
[13] Malthouse, E. C. and Blattberg, R. C. (2005), “Can we predict customer
lifetime value?” Journal of Interactive Marketing, vol. 19, pp. 2-16.
[14] Amin, R. (2011), “using predicted improvement factors for communication
creteria in binary logistic,” operation management.
Vol. 2, No. 1, Ser. 5, (2014), 51-71
Prioritizing Service Organizations Based
on Classified Service Quality Dimensions by
MADM and Importance-Performance Analysis
Reza Dabestani∗
University of Isfahan,
Isfahan, Iran
Arash Shahin
Isfahan, Iran
Hadi Shirouyehzad
Najafabad Branch, Islamic Azad University,
Isfahan, Iran
Mohammad Saljoughian
Isfahan, Iran
Abstract. Current study aims to prioritize four-star hotels through a
two-step procedure: firstly, classifying service quality dimensions (SQDs)
based on Importance-Performance Analysis and secondly prioritizing
hotels based on classified SQDs analysing the results. To reach that aim,
the customers of three 4-star hotels were considered and asked to state
the quality of the service they expected to receive prior to its delivery
as well as their feelings toward it after they received the service. Then,
Importance-Performance Analysis was used to classify SQDs.
∗
51
52
R. Dabestani, A. Shahin, H. Shirouyehzad, and M. Saljoughian
Finally, we exploited TOPSIS and Shannon Entropy to prioritize Hotels. The method taken in this paper, which first categorizes SQDs, is
rather noble. Findings revealed that among the 12 dimensions of service
quality, Competence, Tangibles and Price are the most crucial factors,
and hence should receive more attention in hospitality industry. The results of this research may provide insightful hints to the hotel managers
about those aspects of service that form their customers’ perception of
service quality.
The findings can also help the practitioners to assign resources appropriately and offer a more competitive service to the customers through
paying attention to those factors of service which are of critical importance in this industry.
Keywords: Service quality; prioritization; hotel industry; importanceperformance analysis; TOPSIS.
1.
Introduction
Service quality as an important factor that plays a critical role in the
success of any service organization. Since customers cooperate in delivery services, they interact closely with various aspects of organizations. Therefore, they have the opportunity to evaluate the quality of
services provided by organizations [1]. Customers usually assess service
quality by comparing their perceptions and expectations toward the delivered services [2]. Service quality can influence adding value to the
overall service experience [3].
During the last three decades, the tourism industry has become one
of the most important players of economies worldwide. As the role service industries in modern economies have been more critical, new challenges made in service marketing have received more attention [4]. Competitiveness as a new challenge in each industry is derived from the performance of its enterprises [5]. Service industries must enhance quality of
services that exceed customers’ expectations in order to be successful in
competitive market [6]. Competitive condition in service industries has
forced companies to seek competitive advantages, efficiency and profitable ways in order to get ahead of other firms.
Performance evaluation, as a managerial issue, is not only limited
to some concepts such as productivity, efficiency, etc., and can be anal-
Prioritizing Service Organizations Based ...
53
ysed from different points of view [7]. Mahadeo and Durbarry [8] stated
that companies should deliver appropriate service quality in competitive condition. More recently, a number of researches have been done
to find out the relationships between service quality and organisations’
performance. It has been proved that high quality of service plays a vital
role in the success of organisations [9]. Organisations have realized that
delivering appropriate services can improve financial performance and
customer satisfaction [10]. Chen [11] stated that the performance evaluation factors in hospitality industry are inherently multidimensional
and can hardly be measured. Czepiel [12] stated that business success
depends on the performance of a service provider which is derived from
its interaction with customers. Service quality has been recognized as an
overall evaluation of service by customers. This concept is an indispensable criterion in service evaluation by customers [13].
Since the concept of service quality is inherently intangible, measuring service quality has been a challenging issue [14]. Service quality
in service industries can be described as meeting customers’ needs and
requirements and how well customers’ expectations are fulfilled [9].
A number of scholars and researchers have applied gap concept to
identify critical performance attributes which have the most influence on
customer satisfaction. Deng and Pei [15] stated that due to the existence
of a non-linear relationship between attribute performance and overall
satisfaction, there is a casual relationship between attribute importance
and attribute performance. Therefore, customer’s self-stated importance
may not be the actual importance of service attribute.
This study evaluates hotels based on SQ criteria. To do so, the customers’ perceptions and expectations in three 4-star hotels are measured. Then IPA is used to categorize service aspects based on their
importance. Finally, exploiting Entropy and TOPSIS each dimension of
SQ, each dimension is prioritized and the hotels are ranked.
Service quality is an applicable concept in private sector, because
poor quality of services can negatively influence the reputation of an
organisation. The first step in service quality evaluation is to explain its
definition in order to measure and analyse this concept. Consequently,
it can help service organisations to determine the desired level of quality
54
and its related problems [16]. Service quality is related to different aspects of organisations. Schlesinger and Heskett [17] and Heskett et al. [18]
stated that there are significant relationship between service quality, the
value of services, customer satisfaction, customer loyalty and financial
outcomes of an organisation. Thus, improving the level of service quality can influence the level of customer satisfaction, customer loyalty and
performance of organisations [19].
Service quality has been defined by many researchers and practitioners ([20, 21, 22, 23, and 24]). The common part of all of these studies is
the definition of service quality which is based on the customers’ expectations and perceptions. Lehtinen and Lehtinen [25] introduced physical
and interactive quality while Greenrooms [26] identified three types of
dimensions including technical, functional and firm’s image.
Zeithaml [27] stated that service quality concept is interrelated with
consumer’s judgement about a product’s excellence. However, there is
no consensus about the definition of product quality and its dimensions. Parasuraman et al. [28] proposed ten dimensions for service quality. These dimensions include tangibles, reliability, responsiveness, competence, courtesy, creditability, security, access, and communication and
understanding customer [29].
Parasuraman et al. [22] proposed SERVQUAL approach in which
five dimensions of reliability, responsiveness, tangibles, assurance and
empathy are addressed [30]. Bruck et al. [31] introduced six dimensions
of ease of use, functionality, performance, durability, serviceability and
prestige for durable goods. Shahin [32] proposed a comprehensive list of
SQDs for British Airways and some international and domestic hotels.
In his study, SQDs were classified into 12 major categories in the first
level and 30 items in the second level (Table 1). Comparing the Shahin’s
proposed set of SQDs with other studies, it seems the 12 categories are
relatively more comprehensive and therefore, the authors have decided
to use it for this study.
56
2.
Importance-Performance Analysis
A useful technique to classify various factors of a study into high/low importance category is Importance-Performance Analysis (IPA) [33]. Martilla and James [34] introduced this analysis to identify customer requirements. IPA can provide insightful information on important dimensions
as well as less important ones [35]. The prime advantage of this technique
is the way it presents data, suggests and implies practical, strategic suggestions [36]. These advantages have extended the application of IPA to
a wide range of purposes: it has been exploited as a means for analysing
customer satisfaction [37] tourism management and marketing performance [38], industry [39], banking [40], food industry [41], restaurants
[42], and hotel management [43].
IPA looks like a two-dimensional coordinating system, with its vertical axis usually representing the customer’s satisfaction of a specific
dimension of the service he/she has received, and its horizontal axis
generally reporting the importance of that aspect of service to him/her
[35]. This two-dimensional coordinating system along with its axes is
called IPA grid [44]. IPA grid is further divided into four quadrants,
mostly with arithmetic means of the sample values represented by the
horizontal and vertical axes [33]. The first quadrant-possible overkillholds the aspects the service provider has performed very well, but
maybe too well, because the customer does not care a lot about them. The
second quadrant-keep up the good work-comprises the attributes that
are crucial for the customers, and the service provider has succeeded
to gain their satisfaction in those attributes, and it is important to focus and keep working on them. The third quadrant - concentrate herehighlights the aspects that are crucial for the customer, but the service has failed to comply with his/her expectations. Finally, the fourth
quadrant-low priority-bolds the attributes of the service that neither
satisfied the customer, nor they are important to him/her, which means
the management need not try to improve those aspect even though they
are not performed well.
Using IPA requires a four-step procedure: first, the key factors of the
problem should be recognized. Then, the importance of each factor and
57
also the performance in that factor should be measured. In the third
step, the data should be plotted on the IPA grid by pairing the mean
scores for each attribute measured in step 2. Lastly, the quadrants of IPA
grid should be assigned through what was explained above [45]. Based
on the IPA methodology, the attributes that fall in the first quadrant
are the ones that are receiving too much attention, and hence are wasting the budget; whereas the ones in the second quadrant are managed
efficiently, and the policy should not change for them. The attributes
in the third quadrant need urgent attention, because they are probable sources of customer dissatisfaction; and the attributes in the fourth
quadrant do not need urgent attention, for they are not very important
in the customer’s point of view [46].
A simple yet powerful method used to assign weights to criteriabased on the dispersion and variance-is Shannon Entropy [47].
The process is as follows [48]:
1. Normalizing the criteria
Xij
Pij = j Xij
2. Calculating Ej indicator for each criterion using the formula below
Ej = −k
Pij Ln(Pij )
K=
1
Lnm
m=number of alternatives
3. Calculating Dj indicator
Dj = 1 − Ej
4. Calculating each indicator’s final weight
Dj
Wj = Dj
Technique for order preference by similarity to an ideal solution (TOPSIS) as a multiple criteria method aims to address solutions from a finite
58
set of alternatives [49]. The underlying logic of TOPSIS is to define both
positive and negative ideal solutions [50]. The basic principle is that the
chosen alternative must have the shortest distance from the ideal solution and the farthest distance from the negative-ideal solution. TOPSIS
is a practical and helpful tool for prioritising and selecting the most
appropriate alternative through distance measures [51]. The TOPSIS
procedure consists of the following steps [52]:
Step 1: computing the normalized decision matrix:
xij
nij = m
2
i=1 xij
, nij : the normalized component of the decision matrix
Step 2: Calculating the weighted normalized decision matrix:
V = N D × Wn×n
V: the weighted normalized component of decision matrix
Step 3: Determining the ideal and negative ideal solution:
V + = {Vi+ , Vn+ } = {(maxVij /i ∈ I ), (minVij /i ∈ I )}
V − = {Vi− , ..., Vn− } = {(minVij /i ∈ I ), (maxVij /i ∈ I )}
Where I is associated with the advantage criteria and I is associated
with the cost criteria.
Step 4: Computing the separation measures, utilizing the n-dimensional
Euclidean distance. The distance between each element of the alternative
from that of ideal solution is calculated through the following formulas:
n
Si+ = (Vij − VJ+ )2
i = 1, 2, ..., m
j=1
n
−
Si = (Vij − VJ− )2
j=1
i = 1, 2, ..., m
59
in which Vj+ is the positive ideal option, and Vj− is the negative deal
option.
Step 5: Calculating the relative closeness to the ideal solution. The
relative closeness of the alternative aj with respect to cli+ is defined as:
cli+ =
3.
Si−
Si− + Si+
Research Methodology
Step 1: Measuring customers’ perceptions of Four-Star hotels
In order to measure customers’ perceptions of the service, a five-point
Likert scale was designated based on the 30 items in the second level of
proposed list for SQDs (Table2). In this questionnaire, participants were
asked how they evaluated the performance of the hotels in the delivered
services. In our Likert scale, 1 notified strongly disagree and 5 notified
strongly agree
Step 2: Measuring customers’ Expectations of Four-Star hotels
Another questionnaire was also designed to measure the customers’ expectations of SQDs as brought in Table 2. In this questionnaire, the
customers were demanded to state how important every single SQD in
their points of view is. The second questionnaire resembled the first one
in that it also used a 5-point Likert scale in which 1 notified strongly
disagree and 5 notified strongly agree.
Step 3: Measuring service quality gaps
Service quality gap was calculated from the distance between customers’
expectations and perceptions. This step was completely described in
section three.
Step 4: Importance-Performance Analysis of service quality dimensions
The last stage comprises two steps; firstly, the mean value of expectation and perception should be calculated by simply adding up the scores
of the two and dividing the sum into the number of subjects in each
group. Secondly, the dimensions should be plotted on the IPA grid. The
60
coordination is calculated by simply adding up the perception values
for a certain dimension and dividing the number by the number of participants, then doing the same calculation for the expectation values of
the sub-dimension, and finally pairing the two values to pinpoint the
location of that dimension on the grid.
Step 5: Ranking Hotels by Entropy and TOPSIS techniques
In this step, we utilize gap values as the entry of Entropy and TOPSIS
techniques. Then, we classify the SQDs into three groups of very important (the SQDs in quadrant 3), important (the SQDs in quadrant 2)
and not important (the SQDs in quadrants 1 and 4). Using Entropy we
calculate the weights of SQDs in each quadrant. Finally, we rank the
hotels based on the gap values and the calculated weights by TOPSIS
method.
Hospitality industry highly relies on the customers’ perception from
delivered services. Therefore, service quality, as a critical issue, can play
a vital role in this industry (Lee, 2008). In this study, the questionnaires
were designed based on the 30 sub-dimensions of service quality (second
column in Table 1) and were distributed in three four-star hotels in
Isfahan. There are just four four-star hotels in Isfahan city and only three
of them accepted to cooperate in this study. In these questionnaires,
participants were asked to reflect on their expectations and perceptions
for each sub-dimension of SQ. The sample in each hotel includes 66
Iranian customers.
Aseman Hotel, which is located in the centre of Isfahan City, has 13
floors as well as two quest elevators on each floor and accommodates
customers in 90 rooms. Just about half of the respondents are between
25 to 35 years old (53 percent), 68.2 percent are married and male, which
means only about a third of the customers are single and female. More
than half of the respondents have a bachelor degree (53 percent) and
earn more than 5 hundred dollars per month (75.8 percent).
Ali Qapu Hotel is located in the most ancient neighbourhood of
Isfahan called Chaharbagh-Abbasi. This hotel is close to some historical
places of Isfahan such as Sio-se-Pol Bridge, Naghsh-e-Jahan Square and
Chehel-Sotoun Palace. The majority of participants in this hotel are
61
between 25 to 35 and 45 to 55 years old (51.5 percent), 69.2 percent are
married and 53 percent are female, and 72.7 percent of customers have
higher education degrees.
Piruzi hotel is located in the centre of the city, in Chaharbagh
Street. In this hotel, most participants are young (54.5 percent less than
35 years old). In this group, 10.6 percent are between 15 and 25 years
old and 43.9% are in the category of 25 and 35 years old. More than half
of the respondents (= 56.1%) are male and the rest are female. A great
portion of customers (80.3%) are married and the rest are single. Most
customers (43.9 percent) have four years of academic study and make
more than 500 dollars per month.
4.
Findings
The obtained data passed through the four steps mentioned in section
5. The results are as follows:
Step 1: Measuring customers’ perceptions of Four-Star hotels
As shown in Table 2, customers’ highest and second highest perception in Aseman, Ali Qapu and Piruzi Hotels belong to “reliability” and
“courtesy”, respectively. On the other hand, the customers’ lowest perceptions values in Aseman and Piruzi goes to “price”. However, the
lowest performance value in Ali Qapu Hotel belongs to “tangibles”. A
note worthwhile to mention is that the average value of customers’ perceptions in Ali Qapu Hotel is higher than four.
Step 2: Measuring customers’ expectations of Four-Star hotels
According to Table 3, the customers’ highest expectations in Aseman,
Ali Qapu and Piruzi Hotels refer to “flexibility”, “reliability” and “price”,
respectively. The lowest perception value in Aseman and Piruzi goes to
understanding the customer. However, the lowest performance value of
Ali Qapu Hotel is related to “communication”. It is important to note
that customers of Ali Qapu Hotel, who have greater perceptions, have
greater expectations as well.
64
After calculating the mean values, the dimensions were plotted on
the grid. Figures 1 illustrates the results graphically. As it is shown, one
dimension named “access and approachability” is placed in Q1 (possible
overkill). Six dimensions including “reliability”, “responsiveness”, “creditability”, “flexibility”, “security & confidentiality”, and “courtesy” are
placed in the Q2 (keep up the good work). “Tangibles”, “competence”
and “price” are positioned in Q3 (concentrate here); and “communications”, “understanding the customer”, are plotted in Q4 (Low priority).
Step 5: Ranking Hotels by Entropy and TOPSIS techniques
According to previous step, the SQDs can be categorized into three
groups. The SQDs in the Q3 including “Tangibles”, “competence” and
“price” are considered as the most important criteria. Six dimensions including “reliability”, “responsiveness”, “creditability”, “flexibility”, “security & confidentiality”, and “courtesy” are considered as the important
criteria since they are in keep up the good work quadrant (Q2) and the
SQDs including “access and approachability”, “communications” and
“understanding the customer”, are considered as the less important criteria. In this step, the gap values were fed into the TOPSIS and Entropy
and the process of calculating the weights of criteria and ranking the alternatives are performed for three times. According to the Table 5, the
highest weight is refers to the price and the lowest weight is related to the
tangibles. Also, Ali Qapu and Aseman Hotels are in the first and second
rank, respectively. Considering important criteria in Table 6, creditability and flexibility criteria have the highest weight. Similar to previous
ranking, Ali Qapu Hotel has the best rank. As it is shown in Table 7,
the highest weight refer to the “understanding the customer” criterion
and Ali Qapu Hotel has been determined as the best Hotel.
67
all hotels and as it is clear, all the values are positive. These positive
values point that the performances of hotels are lower than customers’
expectations.
References
[1] Kandampully, J. (2000), The impact of demand fluctuation on the quality
of service: a tourism industry example. Managing Service Quality, Vol.
10, No. 1, pp. 10-19.
[2] Shahin, A. and Dabestani, R. (2010), Correlation Analysis of Service
Quality Gaps in a Four-Star Hotel in Iran. International Business Research, Vol. 3, No. 3, pp. 40-46.
[3] Lau, P. M., Akbar, A. K. h., and Fie, D. Y. G. (2005), Service quality: A
study of the luxury hotels in Malaysia, The Journal of American Academy
of Business, Vol. 7, No. 2, pp. 46-55.
[4] Yu, M. M. And Lee, B. C. Y. (2009), Efficiency and effectiveness of service
business: evidence from international tourist hotels in Taiwan, Tourism
Management, Vol. 30, No. 4, pp. 571-580.
[5] Barros, C. P. (2005), Measuring efficiency in the hotel sector, Annals of
Tourism Research, Vol. 32, No. 2, pp. 456-477.
[6] Gupta, S., McLaughlin, E., and Gomez, M. (2007), Guest satisfaction
and restaurant performance, Cornell Hotel and Restaurant Administration Quarterly, Vol. 48, No. 3, pp. 284-298.
[7] Cook, F. L. (2001), Human resource strategy to improve organizational
performance: a route for films in Britain? International Journal of Management Review, Vol. 3, No. 4, pp. 321-339.
[8] Mahadeo, J. D. and Durbarry, R. (2008), The level of service quality in
the mobile telephony sector: the case of Mauritius, International Journal
of Services and Operations Management, Vol. 4, No. 6, pp. 730-744.
[9] Ramseook-Munhurrun, P., Naidoo, P., and Lukea-Bhiwajee, S. D. (2009),
Employee perceptions of service quality in a call centre’, Managing Service
Quality, Vol. 19, No. 5, pp. 541-557.
68
[10] PalaniNathaRaja, M., Deshmukh, S. G., and Wadhwa, S. (2006), Measuring service quality in technical education and healthcare services, International Journal of Services and Operations Management, Vol. 2, No.
3, pp. 222-236.
[11] Chen, S. H. (2009), ‘Establishment of a performance-evaluation model for
service quality in the banking industry’, The Service Industries Journal,
Vol. 29, No. 2, pp. 235-247.
[12] Czepiel, J. A. (1990), Service encounters and service relationships: implications for research, Journal of Business Research, Vol. 20, No. 1, pp.
13-21.
[13] Ganguli, S. H. and Roy, S. K. (2010), Service quality dimensions of hybrid
services, Managing Service Quality, Vol. 20, No. 5, pp. 404-424.
[14] Kandampully, J. (1997), An exploratory study of service quality in the
Malaysian public service sector, International Journal of Quality & Reliability Management, Vol. 24, No. 2, pp. 170-190.
[15] Deng, W. J. And Pei, W. (2009), Fuzzy neural based importanceperformance analysis for determining critical service attributes, Expert
System with Applications, Vol. 36, No. 2, pp. 3774-3784.
[16] Kakouris, A. P., and Meliou, E. (2011), New public management: promote
the public sector modernization through service quality. Current experiences and future challenges. Public Organization Review, Vol. 11, No. 4,
pp. 351-369.
[17] Schlesinger, L. A. and Heskett, J. L. (1991), Leonard A. Schlesinger and
James L. Heskett respond: customer satisfaction is rooted in employee
satisfaction, Harvard Business Review, Vol. 69, No. 6, pp. 148-150.
[18] Heskett, J. L., Sasser, W. E., and Schlesinger, L. A. (1997), The Service
Profit Chain, 1st ed., the Free Press, New York.
[19] Horn, C. and Rudolf, M. (2011), Service quality in the private banking
business’, Financial Markets and Portfolio Management, Vol. 25, No. 2,
pp. 173-195.
[20] Cronin, J. J. and Taylor, S. A. (1992), Measuring service quality: a reexamination and extension, Journal of Marketing, Vol. 56, No. 3, pp.
55-68.
69
[21] Cronin, J. J. and Taylor, S. A. (1994), SERVPERF versus SERVQUAL:
reconciling performance-based and perceptions-minus-expectations measurement of service quality’, Journal of Marketing, Vol. 58, No. 1, pp.
125-131.
[22] Parasuraman, A., Zeithaml, V. A., and Berry, L. L. (1988), SERVQUAL:
a multiple-item scale for measuring consumer perception of service quality,
Journal of Retailing, Vol. 64, No. 1, pp. 12-40.
[23] Parasuraman, A., Zeithaml, V. A., and Berry, L. L. (1991), Refinement
and reassessment of the SERVQUAL scale, Journal of Retailing, Vol. 67,
No. 4, and pp. 420-450.
[24] Teas, R. K. (1994), Expectations as a comparison standard in measuring
service quality: an assessment of a reassessment, Journal of Marketing,
Vol. 58, No. 1, pp. 132-139.
[25] Lehtinen, U. and Lehtinen, J. R. (1985), Service quality: a study of quality
dimensions, Paper read at the Second World Marketing Congress, University of Stirling, Scotland.
[26] Greenrooms, C. (1984), A service quality model and its marketing implications, European Journal of Marketing, Vol. 18, No. 4, pp. 36-44.
[27] Zeithaml, V. A. (1988), Consumer perceptions of price, quality, and value:
a means-end model and synthesis of evidence, Journal of Marketing, Vol.
52, No. 3, pp. 2-22.
[28] Parasuraman, A., Zeithaml, V. A., and Berry, L. L. (1985), A conceptual
model of service quality and its implications of future research, Journal
of Marketing, Vol. 49, No. 4, and pp. 41-50.
[29] Zeithaml, V. A. and Parasuraman, A. (2004), Service Quality, Marketing
Science Institute, Cambridge, Mass.
[30] Mels, G., Boshoff, Ch., and Nels, D. (1997), The dimensions of service
quality: the original European perspective revisited, The Service Industries Journal, Vol. 17, No. l, and pp. 173-189.
[31] Bruck, M., Zeithaml, V. A., and Naylor, G. (2000), Price and brand name
as indicators of quality dimensions for consumer durables, Journal of the
Academy of Marketing Science, Vol. 28, No. 3, and pp. 359-374.
70
[32] Shahin, A. (2007), In-flight of service quality dimensions: a comprehensive
review, International Journal of Excellence in Tourism, Hospitality and
Catering, Vol. 1, No. 2, pp. 13-27.
[33] Su, Ch. Sh. (2013), An Importance-Performance Analysis of Dining Attributes: A Comparison of Individual and Packaged Tourists in Taiwan.
Asia Pacific Journal of Tourism Research. Vol. 18, No. 6, pp. 573-597.
[34] Martilla, J. A. and James, J. C. (1977), Importance-performance analysis.
Journal of Marketing, Vol. 41, No. 1, pp. 77-79.
[35] Frauman, E. and Banks, S. (2010), Gateway community resident perceptions of tourism development: Incorporating importance-performance
analysis into limits of acceptable change framework. Tourism Management. Vol. 32, No. 1, pp. 128-140.
[36] Chu, R. K. and Choi, T. (2000), An importance-performance analysis of
hotel selection factors in the Hong Kong hotel industry: a comparison of
business and leisure travellers. Tourism Management. Vol. 21, No. 4, pp.
363-377.
[37] Matzler, K., Bailom, F., Hinterhuber, H. H., Renzi, B., and Pichler, J.
(2004), The asymmetric relationship between attribute-level performance
and overall customer satisfaction: A reconsideration of the importance
performance analysis. Industrial Marketing Management, Vol. 33, No. 4,
pp. 271-277.
[38] Griffin, T. and Edwards, D. (2012), Importance-performance analysis as
a diagnostic tool for urban destination managers. Anatolia: An International Journal of Tourism and Hospitality Research. Vol. 23, No. 1, pp.
32-48.
[39] Almanza, B. A., Jaffe, W., and Lin, L. (1994), Use of the service attribute matrix to measure consumer satisfaction. Journal of Hospitality
and Tourism Research. Vol. 17, No. 2, pp. 63-75.
[40] Cunningham, M. A. and Gaeth, G. J. (1989), Using importanceperformance analysis to assess patients’ decisions to seek care in a dental
school clinic. Journal of Dental Education. Vol. 53, No. 10, pp. 584-586.
[41] Jang, S. S., Ha, A., and Silkes, C. A. (2009), Perceived attributes of Asian
onal Journal of Hospitality Management. Vol. 28, No. 1, pp. 63-70.
71
[42] Chen, J. K. and Chen, I. S. (2010), Disparities between services demanded
and services received in Taiwanese restaurants. Global Journal of Business
Research. Vol. 4, No. 1, pp. 59-69.
[43] Mohsin, A. and Lockyer, T. (2010), Customer perception of service quality
in luxury in New Delhi, India: An exploratory study. International Journal
of Contemporary Hospitality Management. Vol. 22, No. 2, pp. 160-173.
[44] Oh, H. (2001), Revisiting importance-performance analysis. Tourism
Management. Vol. 22, No. 6, pp. 617-627.
[45] Lai, L. S. L. and To, W. M. (2010), Importance-performance analysis
for public management decision making: An empirical study of China’s
Macao special administrative region. Management Decision. Vol. 48, No.
2, pp. 277-295.
[46] Abalo, J., Varela, J., and Manzano, V. (2007), Importance values for
importance-performance analysis: A formula for spreading out values from
preference rankings. Journal of Business Research, Vol. 60, No. 2, pp. 115121.
[47] Zeleny, M. (1996), Multiple criteria decision making, New York: Springer
[48] Wang, T. and Lee, H. (2009), Developing a Fuzzy TOPSIS Approach
Based on Subjective Weights and Objective Weights. Expert Systems with
Applications. Vol. 36, No. 5, pp. 8980-8985.
[49] Chen, S. J. and Hwang, C. L. (1992), Fuzzy Multiple Attribute Decision
Making: Methods and Applications, Springer-Verlag, Berlin.
[50] Hwang, C. L. and Yoon, K. (1981), Multiple Attribute Decision Making:
Methods and Applications, A State of the Art Survey. Sprinnger-Verlag,
New York, NY.
[51] Tan, Ch. (2011), a multi-criteria interval-valued intuitionistic fuzzy group
decision making with Choquet integral-based TOPSIS, Expert Systems
with Applications, Vol. 38, No. 4, pp. 3023-3033.
[52] Opricovic, S. and Tzeng, G. H. (2004), Compromise solution by MCDM
methods: A comparative analysis of VIKOR and TOPSIS, European Journal of Operational Research, Vol. 156, No. 2, pp. 445-455.
Vol. 2, No. 1, Ser. 5, (2014), 73-89
A 360 Degree Performance Appraisal
Model for Documents Digitizing Firms
Amir Hossein Koofigar∗
Isfahan Payame Noor University,
Isfahan, Iran
Mahshid Ghaziasgar
Isfahan Payame Noor University,
Isfahan, Iran
Mahdi Karbasian
Malek Ashtar University,
Tehran, Iran
[email protected]
Abstract. Performance appraisal is a process that people can compare their perceptions of working with major manufacturers’ perceptions. However, when assessment is done from different sources that
are related with self-assessment, called performance appraisal 360 degree. Given the importance of Performance appraisal by 360 degree
feedback in organizations, a model has been provided for 360 degree
performance appraisal of firms that digitize documents. The model was
confirmed 40 experts by questionnaire method, Also reliability of the
questionnaire was confirmed by Cronbach’s Alpha. The model was implemented in a similar firm. By determining of expectations and criteria
of performance evaluation of each sector the other sectors, The final 17
different performance evaluation forms were developed that will be completed by its members and customers to performance evaluation of the
other employees. The results of data and review it and recommendations
were provided, improve the productivity of the organization.
Keywords: Performance appraisal; 360 degree feedback; performance
evaluation criteria; proposed model.
∗
Received: January (2014); Final Revision: April (2014)
73
74
1.
A. H. Koofigar, M. Ghaziasgar, and M. Karbasian
Introduction
In the present day competitive world only the organizations can remain
in this world that uses their sources in the best manner. One of the
important organizational sources is man power. With regard to the case
that the employees need to know the organization’s expectations from
themselves and the rate of their appraisal for removing their past deficiencies and also promoting effectiveness and efficiency as well as finding
their abilities, appraisal of man power performance is a very important
procedure and one of the most sensitive problems for the authorities in
the organization. In many of the organizations the employees are not informed sufficiently for their performance manner and they do not know
that their performance is desired or undesired, or whether there is an
improvement in their performance. In fact, the employees believe that
the directors do not wish to satisfy these needs that today it is one of
the great challenges for the organizations. Thus the directors shall give
feedback to the employees who have better performance, and not to
waste most of their time for improving the behavior of the troublesome
employees [12].
The system of performance appraisal is the process of recognition,
evaluation and development of the individuals’ performance for achieving to the individual and organizational purposes [8]. And it helps the
individuals to compare their comprehension in their working environment with the comprehension of the important appraisers. Performance
appraisal is analyzing the successes and failure of an individual and
studying his/her competencies for the future job training and promotion [15]. In other words, performance appraisal is for this case that
what works shall be performed in which places in order to become more
successful, the final purpose of performance appraisal is increasing effectiveness and efficiency. Today the organizations have found that performance appraisal system has abundant power to the extent that it can
change the organization’s culture [1].
In an organization, performance appraisal is a way that through
which the employees’ performance information is obtained for important
decision makings such as: salary and wage, propaganda, recognition of
A 360 Degree Performance Appraisal Model ...
75
training and development needs via the performance level or their behaviors. In addition an evident relation is observed between performance
appraisal and the employees’ approaches, behaviors and efforts that indicate the improvement of obtained financial results by the organization
[10].
In order that the directors make the employees aware of their performance manner, they shall become familiarized with the newest feedback skills and create an appropriate method for presenting the continuous feedback in the organization[21] that consist of some methods such
as: privileging method, obligatory selection method, method of registering sensitive events, the method based on purposeful management,
degree method, individual to individual comparison method, feedback
method, that the feedback method itself consists of 180, 360, 540 & 720
degree feedback.
In a research it was determined that in case that the appraisers are
unknown, the appraisal will be performed more real. When the appraisers had been known, they liked to perform a positive appraisal and regard the realities in their appraisal less. This weakness which mostly was
observed in the 180 degree appraisals became a base for the research and
creation of more challenging models such as 360 degree feedback that
plays important role in the process of obtaining organization’s feedback
[17].
The process of performance appraisal is based on the views of various appraisal groups which are in relation with the appraised employees,
in fact they express that how the employees can develop and improve
their job. This process, including each individual’s view is about him/
her as well. This type of performance appraisal is called 360 degree
appraisal or inseparable appraisal [9] that dominates on some of the
defects of traditional appraisal such as non-objectivity, bias or halo errors [6,10]. Other common terms that are used for 360 degree feedback
consist of: “Beneficiaries appraisal”, “multi-criteria feedback”, “Multisource appraisal”, “subordinates appraisal”, “Group appraisal”, “multilateral or multi-degree appraisal” [14].
Also the definition of Ward [20] from 360 degree appraisal is this: Regular compilation of a group or an individual’s performance feedback data
77
2. To select data collection tools: There are some questionnaires that
are completed by the appraisers, some of the organizations used interview as well.
3. To make decision in some cases: In this field, it is recommended
that real behavior of the individual to be considered instead of his/her
general characteristics. The behaviors that are appraised shall be resulted from the organization’s perspective and values.
4. To make decision concerning feedback receivers: In this phase feedback receivers are determined. In the managerial literature, this belief
is mentioned that the individuals shall be volunteered for participating
in the feedback plan. Compulsory participation can endanger system
effectiveness.
5. To train the appraisers and appraised people: training the appraised people is essential in the field of accepting negative feedback. The
appraisers shall become aware in the field of various appraisal errors that
may occur.
6. Feedback receivers select the appraisers: This is one of the literary
fields which are argumentative due to this reason that some of the writers
feel that the receivers may select some of the appraisers in their appraisal
who create easier atmosphere.
7. To distribute questionnaire: The questionnaire has two forms. One
method is paper and pen format and the other method is sending disk
to each one of the appraisers.
8. To analyze feedback information: In this stage feedback information are collected and necessary reports are provided.
9. To feedback the feedback: Once the reports were made and the
final report was completed, the feedback is presented to the feedback
receivers.
10. To follow up the process execution: The feedback receivers shall
have practical plan in the field of removing their weaknesses.
11. To repeat the process: In the organizations that 360 degree feed-
78
back works effectively, it is possible that the process to be repeated after
receiving the initial feedback reports [14].
Prevalent Models in 360 degree feedback:
1. The first model is “job analysis”. This type of appraisal assesses
knowledge, skills and abilities based on the procedures for the traditional
job analysis.
2. The second model is based on the individual capability. Instead
of simple measurement of the skills and abilities, it is concentrated on
the appraisal of the capabilities related to the special job.
3. The third model is concentrated on the strategic planning. This
type of 360 degree appraisal assesses knowledge, skills and abilities based
on the organization’s strategic plans. This is believed that these behaviors or abilities help the organization to achieve to its strategic aims and
plans.
4. The fourth model is resulted from development theory that is in
close relation with theoretical and conceptual growth and development
models by Karraheh. On this basis that 360 degree feedback is resulted
in increasing consciousness and in fact to the more effective Karraheh
development processes.
5. The fifth model is based on the character theory. This model assesses some skills, knowledge, and abilities which are in relation with the
character (such as qualities, characteristics, specifications, communication styles, interpersonal relations and individuals’ recognition) [14].
2.
Literature Review
Azar & Sepehrirad have presented mathematical model for 360 degree
performance appraisal. They express that developing a comprehensive
model is essential that weighs and collects the mental judgments for
various appraisal sources. At first they grouped performance appraisal
indices in the four groups of individual characteristics, technical skills,
human skills, and comprehensive skills, and each one was given a weight,
79
then a weight was considered for each one of the appraisal sources using
phase AHP Technique (views by 3 experts). For determining the final
point for the performance of each employee in this research the suggested
model by Anders has been used [19].
In this direction, a model for appraising 360 degree performance
with the title of heterogeneous information and affiliated criteria has
been presented. With expressing that various criteria exist in appraising
performance that may have different nature or non-exactness in present,
presenting a heterogeneous framework for these criteria has been considered essential. In this field, the criteria are appraised with regard to
the rate of appraisers’ information from the employees under study. Also
an integrated model has been suggested for 360 degree appraising that
makes possible an inflexible appraisal framework and the arbitrators can
present their appraisals in various fields with regard to non-exactness
and nature of the criteria. For getting assurance from effectiveness of
the collected information, in this model a set of the factors for controlling the criteria effect and the rate of arbitrators’ effect has been created
[9].
Also a research has been performed under the title of “Relation between Appraiser’s Effect and three Source for 360 Degree Feedback Appraisal” that studies if the appraiser’s effect has a similar effect on easy
appraisals of the three sources (superior, subordinate and colleague) and
also is there any interaction between the appraiser’s effect and the time
that he spends for observing the appraised individual. The obtained results indicate the influence of appraiser’s effect on easiness of superior,
and colleague appraisals more than the subordinates’ feedback and also
it indicates that this effect increases with increasing the observance time
[5].
From other performed researches we can refer to a research with
the purpose of studying the sexuality effect in managers’ performance
appraisal with execution of 360 degree appraisal in the International Organization of England Financial Services, in this research, it is expressed
that with regard to the essential role of the performance appraisal, from
one side it confronts with a great pressure concerning non-influencing
the appraisals against illegal discriminatory variables like: age, race and
80
sexuality. In general, with performed researches there is no considerable
deed expressing statistical sexual bias from obtained appraisals results
for women managers from the 4 sources: self-appraisal, supervisors’ appraisal, subordinates’ appraisal, and colleague’s appraisal. In fact, after
collecting the appraisals results from all of these groups, the final point
given to the women managers’ performance was considerably higher than
their men counterparts. And this is in contrast with the traditional viewpoint of sexual discrimination in which women are prone to negative view
based on sexuality [16].
Also a research has been performed in the field of comparing the
results of performance appraisal with the traditional method and 360
degree feedback and its relation with satisfaction of the employees at
the hospitals in Lorestan University of Medical Sciences, and according
to the obtained results, there is considerable difference between the appraisal results with the traditional method and 360 degree feedback. Appraisal with the method of 360 degree feedback had more effect on the
employees’ satisfaction; also there is significant and reverse difference
between traditional appraisal and employees’ satisfaction and it means
that the existing method has reverse relation with employees’ satisfaction and the more this method is used, the fewer employees’ satisfaction
is created [13].
Another research in this field with studying the effect for agreement
or disagreement of student sexuality as appraiser and teacher as the
appraised one expresses that with regard to the obtained results from
performed researches in Mashhad Ferdowsi University agreement or disagreement of sexuality has no effect on the appraisal of the students
from the scientific board members [2].
Nelson expresses that in 1994, 22 companies out of 32 famous companies of Fortune Magazine used 360 degree feedback (down up) in a
manner that using 360 degree feedback was public and general almost
among company of Fortune Magazine [18].
3.
81
Main Body
In the previous sections 360 degree performance appraisal was discussed.
In this section with consideration of the importance of 360 degree performance appraisal in the organizations, a model is presented for 360 degree
performance appraisal in the deeds digitalizing services companies and
then it is analyzed.
As it is observed in figure 2, the presented model is for the deeds
digitalizing companies that such companies while participating and admitting in the desiring organ bid, receive the archived deeds of the individuals in that organization and change these documents and deeds
from paper to digital, also with designing and presenting software for using the electronic deeds complete the project. In this manner, following
sections exist with their related duties for performing the project.
Manager: With participation in the bid and admittance and concluding contract with the related organization, he/ she delivers the related
documents and deeds, also presents a design for designing a software of
using electronic deeds and approves the software. Also the manager is
in charge of the required facilities and employees’ salary.
Technical Manager: He/she related the manager with the supervisors
of each section, and also monitors the supervisors’ work.
Computer director: He/she is in charge of supervising computer employees for creating the related software.
Computer employees: Producing the designed software.
Digital director: He/she is in charge of educating and supervising
digital employees, and also encoding for separating the relative deeds
for each person.
Digital employees: They are in charge of clarifying the deeds and
their digitalization.
Deeds director: He/she is in charge of educating and supervising
deeds employees.
Deeds employees: Registration of electronic deeds.
84
tion.
Appraisal of the colleagues including: computer supervisor-deeds supervisor (R10), Digital supervisor-deeds supervisor (R9), computer supervisor-Digital supervisor (R11), deeds employees with other employees
in their section (R22) and other sections (R18, R19), computer employees with other employees in their section (R23)-and other sections (R19,
R20), digital employees with other employees in their section (R21) and
other sections (R18, R20): Observance of working relations and mutual
respect, cooperation and aid spirit.
As it was observed in the manager’s self-appraisal, self-appraisal criteria in each section are a set of expected criteria of other sections from
the section itself and self- appraisals have not been entitled.
As it was stated, model elements is relation of each section with
other sections and the section itself that in fact express relation of each
section with itself and other sections. With regard to the duties and
responsibilities of each working section, some criteria and specifications
are considered for appraisal that non-existence of these criteria is threatening for the organization’s working system. For example, the duty for
the section of deeds employees is naming the electronically deeds, so
exactness and speed in the performance of this section’s employees is
important that its appraisal is duty of their supervisor, it means the
deeds supervisor. For the other sections also the same action is done.
4.
Model Analysis
After designing the model, a questionnaire was prepared in Likert Standard for approving and efficiency of the presented model in order to
appraise 360 degree performance of deeds digitalizing companies and
in this way approval of 40 experts obtained. For determining the questionnaire reliability, the method of Cronbach’s alpha was used. If the
amount of Cronbach’s alpha in the questionnaire to be more than 0.7,
that questionnaire has good reliability [11]. Also the software SPSS 20
has been used for statistical analyses.
Cronbach’s alpha of this questionnaire is 0.872 and indicates that it
has appropriate reliability.
85
After that, with consideration of the model elements that represent
the expectations and criteria for the performance appraisal of each section compared to the other section, one form of performance appraisal
has been provided for each one, and with regard to the uniformity of the
appraisal criteria for some of the sections compared to the other sections,
after collecting and adjusting the forms, finally 17 performance appraisal
forms were obtained that the content of each form differed depending
on the appraised criteria. These forms are used by the organization’s
members and clients for performance appraisal.
For example in appraisal of the manager, 3 forms are used. One form
for appraisal of the clients from manager, one form for the manager’s
self-appraisal and the last for is for appraisal of the supervisors and technical manager from manager. In another appraisal of the organization
members also the appraisal forms are used in this manner.
The criteria under appraisal which were explained in the previous
section are divided in general in the form of Table 1 that this appraisal
is performed through the provided forms for the organization.
In each form some questions have been determined for appraisal of
each criterio and Likert Scale has been used for replying to each question. With implementing the model and studying the obtained information and results from the performed appraisals, some suggestions were
presented for improving the employees’ performance which are observed
in Table 2.
With regard to the Table 2, remuneration and bonus have been considered for the scores (80-100) that promotion is performed if possible.
Concerning the training for the scores (60-80) and (40-60), at first
the appraisal forms are referred to and lowness of the performance is
studied that in which one of the appraisal criteria it has had weakness
and necessary teachings shall be performed in that field.
And finally for the scores of less than 40, more salary will be reduced
compared to the scores of (40-60) and necessary teachings will be performed with studying observed weaknesses, with this difference that the
obliged appraised person will be dismissed in case that in the later appraisal after completion of the training course do not obtain admittance
mark.
86
5.
Conclusion and Suggestions
Performance appraisal plays a key role in the organization that with
regard to the obtained information important decisions are made regarding the organization, and each individual can improve and develop
his/her job in this manner; while these appraisals are collected from several various sources which are involved with the appraised person that
consist of the individual, then the performance appraisal is called 360
degree.
In this paper, 360 degree performance appraisal model has been designed and implemented for the digitalizing services companies that at
first for effectiveness of the presented model, a questionnaire was provided in Likert Scale and through which 40 individuals of the experts
approved the presented model, and reliability of the questionnaire itself
was confirmed by Cronbach’s alpha.
Then the model elements were determined and a form was provided
for each one with consideration of the appraisal criteria that finally they
were summarized in 17 forms for 360 degree performance appraisal that
these forms were completed by the organization’s members and clients
for appraising the organization’s members and desired results obtained
after deriving information and applying suggestions.
Since execution of each plan and new model in an organization is
confronted with some problems and obstacles, execution of this design
is not an exception to this rule. Non-awareness of the organization’s
members causes that they do not wish to attend in the appraisal as the
appraised and appraiser individuals. Since attendance in this plan shall
be completely voluntarily and each type of compulsion is threatening, for
removing this problem, the management shall in some sessions clarifies
the purpose of appraisal for the organization’s members very well that
to what extent does the design has positive effect in the output of the
members work and in the whole organization and for attracting reliance
of the members it shall be expressed that at first the results for appraisal
of each individual will remain secret and only it will be declared to the
person himself/herself.
Secondly, in case of any weakness in work their working situation
88
References
[1] Abaspour, A. (1384), Manager Technical Manager Supervisors Employees.
[2] Ahanchian, M. R. (1382), Opposing effects of gender on student evaluations of faculty members. Journal of Psychology and Educational Sciences,
183-199.
[3] Andres, R., Garcia-Lapresta, J. L., and Martinez, L. (2010), A multigranular linguistic model for management decision-making in performance
appraisal. Soft Computing, 14 (1), 21-34.
[4] Andres, R., Espinilla, M., and Martinez, L. (2010), An extended hierarchical linguistic model for managing integral evaluation. International
Journal of Computational Intelligence Systems, 3 (4), 486-500.
[5] Antonioni, D. and Park, H. (2001), The relationship between rater affect
and three sources of 360-degree feedback ratings. Journal of Management,
27, 479-495.
[6] Banks, C. G. and Roberson, L. (1985), Performance appraisers as test
developers. Academy of Management Review, 10, 128-142.
[7] Chang, G. (1381), 360-Degree Method to evaluate the performance of
service firms. (Translator ashkzari, J).
[8] Dessler, G. (2000), Human Resource Management. 8th ed., New Jersey:
Prentice.
[9] Espinilla, M., Andres, R., Martinez, F. G., and Martinez, L. (2013), “A
360-degree performance appraisal model dealing with heterogeneous information and dependent criteria”. Information Sciences, 222, 459-471.
[10] Fisher, C., Schoenfeldt, L. F., and Shaw, J. B. (2006), Human resources
management. Boston: Houghton Mifflin Company.
[11] Hair, J. F., Black, W. C., Babin, B. J., and Anderson, R. E. (2009),
Multivariate Data Analysis. 7th Edition, Prentice Hall.
[12] Hinkin, T. and schriesheim, C. (2004), “If you don’t hear from me you
know you are doing fine”. Cornell Hotel and Restaurant Administration
Quarterly, Vol. 45.
89
[13] Javaherizadeh, N., Mehrabi, J., and Bazvand, F. (1390), Compare the
traditional method of performance evaluation and its relationship with
360 degree feedback and employee satisfaction, Lorestan University of
Medical Sciences. Journal of Development, 57-67.
[14] McCarty, M. and Caravan, T. N. (2001), “360 Feedback and processes:
Performance improvement and employee career development”. Journal of
European Industrial Training.
[15] Maund, L. (2001), Introduction to Human Resource Management.
[16] Millmore, M. and Biggs, D. (2007), “Gender differences within 360degree managerial performance appraisals”. The current issue and full
text archive of this journal is available at: www.emeraldinsight.com/09649425.htm.
[17] Mum Ford, A. and Gold, J. (2009), “Management Development Strategies for Action”. Chartered Institute of Personal and Development Cipd
house”. Camp Road, London, Sw 194 ux.
[18] Nelson, B. (2000), “Performance management. The use of informal rewards in recognizing”. Performance; U.S.A. Available at:
http://www.pmanagement.com.articles/9902b.htm
[19] Sepehrirad, R., Azar, A., and Sadeghi, A. (2012), “Developing a hybrid
mathematical model for 360-degree”. Procedia-Social and Behavioral Sciences, 62, 844-848.
[20] Ward, P. (2003), “360 Degree feedback”. The Cromwell press.
[21] Weiss, G. (2004), “How to give and receive employee feedback: Tell staffers
early and often how they performing, and listen to their job-related concerns”. Medical Economical, Vol. 81.
Vol. 2, No. 1, Ser. 5, (2014), 91-103
Application of Genetic Algorithm in
Development of Bankruptcy Predication
Theory Case Study: Companies Listed on
Tehran Stock Exchange
Mohsen Hajiamiri∗
Zahedan Branch, Islamic Azad University,
Zahedan, Iran
Mohammad Reza Shahraki
University of Sistan and Baluchistan,
Zahedan, Iran
Seyyed Masoud Barakati
University of Sistan and Baluchistan,
Zahedan, Iran
[email protected]
Abstract. The bankruptcy prediction models have long been proposed as a key subject in finance. The present study, therefore, makes an
effort to examine the corporate bankruptcy prediction through employment of the genetic algorithm model. Furthermore, it attempts to evaluate the strategies to overcome the drawbacks of ordinary methods for
bankruptcy prediction through application of genetic algorithms. The
sample under investigation in this research includes 70 pairs of bankrupt
and non-bankrupt companies during 2001-2011. Having examined the
obtained data from financial statements of the companies under study,
5 financial independent variables were identified so as to be used in the
model. The results indicated that employment of genetic algorithm in
∗
91
92
M. Hajiamiri, M. R. Shahraki, and S. M. Barakati
predicting financial bankruptcy is highly effective, to the extent it managed to correctly predict the financial bankruptcy of companies two
years before the base year, one year before the base year and the base
year at accuracies of 96.44, 97.94 and 95.53, respectively.
Keywords: Bankruptcy; bankruptcy prediction; multiple discriminant
analysis; logistic regression; neural networks; genetic algorithm.
1.
Introduction
Prediction is important in many aspects of life, i.e. any weak prediction
can lead to inefficient decisions. In fact, any planning, decision-making
and other key tasks associated with managers would face failure without
proper predictions. Generally, the purpose of prediction is to reduce risk
in decision-making. Since prediction cannot be completely eliminated,
it is essential for the decision-making process to explicitly consider the
results of remaining uncertainties in prediction [1].
From the perspective of macroeconomic theories, the level of economic development of a society correlates with the level of investments
made in it. If such investments are not made in the right opportunities
or directed in inefficient ways, the national economy will be damaged
[14].
One of the strategies to assist investors is to offer prediction models about the financial status of companies. The closer predictions are
to reality, the more appropriate the basis of decisions become. The
bankruptcy prediction models are regarded as a tool for estimating the
future performance of companies. Investors and creditors extremely tend
to predict the bankruptcy of businesses or else, great costs are imposed
on them. There are advantages and disadvantages to employment of each
prediction model [3].
Selection of a model according to the consumer needs for financial
information and their environmental circumstances is complicated. In
fact, national wealth can be preserved in the form of physical and human
capital if the probability of corporate bankruptcy in businesses is correctly predicted and the corporate affairs are adjusted through detecting
the problems to be solved. Furthermore, such model can provide an ideal
Application of Genetic Algorithm in Development ...
93
guideline for financial decision-makers, i.e. investment firms, banks and
government [8].
2.
Problem Statement
The growing competition among businesses has restricted the chances
of gaining profits while making bankruptcy more probable. Decisionmaking on financial issues has always involved risk and uncertainty. One
of the strategies to assist investors is to offer prediction models about
the overall outlook of companies. The closer predictions are to reality,
the more appropriate the basis of decisions become. As Beaver argues:
“Prediction is possible without decision-making. The smallest decisions,
however, cannot be made without prediction” (Beaver, 1996). As one of
the strategies for predicting the future status of companies, a bankruptcy
prediction model estimates the probability of bankruptcy through combining a group of financial ratios. Being a telltale sign of misallocation of
resources, the ability to financially and commercially predict is considered vital from the perspective of private investors and also from social
perspective. The early warning of possible bankruptcy enables managers
and investors to take preventive measures and distinguish desirable investment opportunities from those undesirable [11].
In Tehran Stock Exchange, the criterion for bankruptcy and removal
of companies from the list is Article 141 of the Iranian commercial code,
which states: “If at least half of the capital of a company is lost due
to the caused damages, the board of directors shall immediately call on
the stockholders for an extraordinary general meeting in order to discuss
whether the company should be dissolved or continue operating. In case
the board does not vote for dissolution, the corporate capital shall, at
the same meeting under Article 6 of the mentioned law, be reduced to
sum of the currently available capital. In case the board of directors,
against the provisions of this Article refuses to call for the extraordinary
general meeting or the members invited fail to gather, the stakeholders
can individually request the dissolution from a competent court of law”.
[6].
95
model was 94% evaluating 65 different financial ratios in the previous
studies [17].
It should be noted, however, additional studies have been done for
comparing various neural networks. In 2010, for instance, Khashman A.
compared different neural networks in order to predict credit risk. In
his research, Khashman put different structures of neural networks into
comparison [7].
One technique employed in analysis of financial crisis is genetic algorithm. In 1998, Varetto was the first scholar strictly employing the genetic algorithm for bankruptcy prediction. The sample in his study consisted of 500 firms; a total of 236 bankrupted and 264 non-bankrupted
firms. The results suggested an accuracy of 93% for one year before
bankruptcy and 91.6% for two years before bankruptcy [16].
In 2006, Mein et al simultaneously employed the genetic algorithm
and support vector machine, dubbing it GA-SVM. The results of their
study indicated an accuracy of 86.53% for the training set one year
before bankruptcy [12].
In a 2006 study using genetic programming, Lensberg identified 6
out of 28 potential variables of bankruptcy previously examined as significant [9].
4.
Methodology
In the present study, the genetic algorithm model was employed so as to
offer a solution to tackle weaknesses of ordinary methods for bankruptcy
prediction. In terms of methodology, therefore, it is a mathematicalanalytic research. Moreover, it can be regarded as a case study in terms
of research type and developmental in terms of objective.
There are several fundamental steps taken in conducting the present
study as below:
1-Identification of financial ratios in order to predict bankruptcy.
2-Calculation of financial ratios and other required parameters as
independent variables used in the tested model.
3-Classification of firms into bankrupted and non-bankrupted under
Article 141 of the Iranian commercial Code.
96
4-Evaluation of accuracy in the genetic algorithm prediction model
for bankruptcy prediction. The statistical population in this study includes the entire companies listed on Tehran Stock Exchange (TSE)
during 2001-2011. The quality and accessibility of information regarding financial statements were two factors contributing to selection of
such a population.
The statistical population was divided into two categories, the first of
which covers bankrupted companies. The criterion taken into account for
bankruptcy was Article 141 of the Iranian Commercial code. The second
category consisted of survived company’s not encountered bankruptcy. For
data collection, effort was made to select non-bankrupted companies
similar in terms of industry and size, except for cases it was impossible due to extremely small industry size, which can be regarded as one
of the study’s restrictions. Since financial information used regarding
each company covers two years prior to bankruptcy, it can generally
be stated that corporate information between 1999 and 2009 has been
employed. The base year (t) regarding bankrupted companies refers to
the year at which a company faces financial crisis or bankruptcy. With
regard to non-bankrupted companies, the base year refers to the year at
which information from two previous years have been collected.
Throughout the examinations done in the present study, a total of
82 companies over the defined period were subject to the mentioned
law. A few of companies, however, were different from other samples in
terms of financial ratios, which led to poor performance and accuracy
of prediction models. Consequently, a number of bankrupted companies
and their selected pairs were removed from the research reducing the
count of remaining pairs used in the study to 70 (i.e. 70 bankrupted
companies and 70 non-bankrupted companies).
From a systematic viewpoint, it is highly essential to have appropriately valid inputs in order to achieve the right outputs. Since accurate and correct information was required more than anything else, the
data regarding the tested financial ratios were obtained from the public
archive of TSE financial statements in CD form as well as a software
called Rahavard-e-Novin.
In order to refine the collected data from financial statements of the
97
sample companies, MS Excel was employed. Then SPS was used to statistically analyze the refined information. Furthermore, MATLAB was
employed since there were non-linear relationships among the financial
data and the objective was to predict the bankruptcy of companies listed
on TSE.
Since these financial ratios have been widely employed in previous
studies, a total of 15 basic ratios as telltale signs of bankruptcy in a
few previously proposed models for financial bankruptcy prediction were
selected taking into account the fact that ratios were gathered from every
major analytical perspective such as liquidity, profitability, liquidation,
etc.
The initial analysis of variable was conducted through 7 computer
operations on genetic programming algorithm. For each operation, the
results of variable were reported after every 4000 periods. This led to
50 reports which consisted of 15 variables examined in order to determine whether or not it contribute to classification capability of the best
program at the operation time. If the variable left non-zero impact on
the classification capability of the best program, 1 value was added to
it. In other conditions, it received 0 values.
Finally, six financial ratios were used as independent variables enumerated below:
1) Immediate ratio, i.e. immediate assets divided by current debt.
2) Debt ratio, i.e. total debts divided by total assets.
3) Return on assets ratio, i.e. net profit divided by total assets.
4) Profit to revenue ratio, i.e. profit divided by total income.
5) Gross profit ratio, i.e. gross profit divided by total income.
6) Shareholder’s return on equity, i.e. net profit divided by shareholder’s equity.
The genetic programming algorithm is a technique allowing the researcher to find a solution to problem without the need to predetermine
the model. It implies that solution can be any model mathematically
describable. The purpose is to allow the data to as much as possible
represent the facts, so that minimize the level of previous structure offering functional forms and statistical methods of selection [9].
Basically, the genetic programming algorithm is supposed to take the
99
In other words, if the evaluation of genetic programming tree leads
to a numerical value higher than zero, then the examined company falls
under the category of companies progressing toward bankruptcy. If the
value is lower than or equal to zero, the company falls under the category
of profitable.
Since these financial ratios have been widely employed in the employed database might be extremely imbalanced that only 5 to 6% of
the available companies are bankrupted, which should be taken into account in order to design the fitness function. Otherwise, the assessment
might turn into a convergent structure sorting the entire companies as
profitable. In fact, they are not sorted from the first place and the obtained success rate becomes highly favorable. There are three ways to
tackle such issue:
• Under sampling of the larger set
• Oversampling the smaller set
• Change in value (weight) regarding missing of the positive and negative
set to compensate for the imbalanced ratio. For instance, if the imbalanced ratio of 1/10 is in favor of the negative set, then the outcome of
sorting the positive sample should be 10 times greater [2].
Therefore, the fitness function can be formulated as below:
n
f itness =
ui
i=1
where u=
0: incorrect sorting
1: Correctly sorted bankrupted companies
nb =0
nb =1 :
Correctly sorted profitable companies
nb = 0 Is the number of bankrupted companies in the training set, while
is the number of profitable companies in the training set.
Table 1 illustrates the major parameters taken into account for assessment.
6.
101
Discussion and Conclusion
Financial bankruptcy is a crucial issue affecting the economies throughout the world. The extravagant social costs suffered by various stakeholders in connection with bankrupted companies leads to an inquiry
for empowerment of prediction and better understanding of this theory.
The major problems tackled in the present study were imbalance
between the number of companies progressing toward bankruptcy and
the number of profitable companies as well as the amount of unavailable
information in the database used for analysis. The approach adopted
for solving this problem was normalization of data and employment of
a fitting function solving the imbalance problem. The obtained results
were highly favorable. As it was mentioned earlier, the best GP structure achieved successful percentages of approximately 99.7 and 97 in the
training set and the testing set, respectively.
The results obtained from examining each year has been shown in
Table 2, the first row of which indicates the results of two years before the
base year, the second row indicates the results of one year before the base
year, and finally the third row indicates the results of the base year. Each
table illustrates the obtained results from training, testing and the combination. The first column shows the percentage of achieved successes
(i.e. number of correct predictions), the second column shows the percentage of true positives (TP, i.e. the number of companies correctly
sorted as bankrupted), and the third column shows the true negatives
(TN, i.e. the number of companies correctly sorted as non-bankrupted).
References
[1] Kurdestani, A. M. (1996-97), “Profitability used for predicting the cash
flow and future profits”, Journal of Accounting and Auditing Reviews 18
& 19, P42-55.
[2] Alfaro, E. and Sharman, k. (2007), “A Genetic Programming Approach for
Bankruptcy Prediction Using a Highly Unbalanced Database”. European
Journal of Evolutionary Computing, 93, 132-143.
102
[3] Altman, E. (2000), “Predicting Financial Distress of Companies”. Retrieved on September 4th, working paper.
[4] Beaver, W. (1996), “Financial Ratios as Predictors of Failure”. Journal of
Accounting Research, 666-16.
[5] Dimitras, A., Zanakis, S., and Zopudinis, C. (1996), “A survey of business
failures with an emphasis on failure prediction methods and industrial
applications”. European Journal of Operational Research, 90 (3), 487513.
[6] Jahangir, M. (2000), “Commercial Code with Cheques act. The amended
registration regulation of non-commercial organizations”, Tehran, Didar
Publications.
[7] Khashman, A. (2010), “Neural networks for credit risk evaluation: Investigation of different neural models and learning schemes”. Expert Systems
with Applications, (37), 6233-6239.
[8] Lee, K. C., Han, I., and Kwon, Y. (1996),“Hybrid neural network models
for bankruptcy predictions, Decision Support Systems”, (18), 63-72.
[9] Lensberg, T., Eilifsen, A., and McKee, T. E. (2006), “Bankruptcy theory
development and classification via genetic program”. European Journal of
operational research, 169, 677-697.
[10] Martin, D. (1977), “Early warning of bank failures: A logit regression
approach”. Journal of Banking and Finance, 1, 249-276.
[11] Mehrani S., Bahramfar, N., and Ghayur, F. (2005), “A Study on the
Correlation between the Traditional Liquidity Ratios and Ratios of Cash
Flow Statement for Assessing the Continuity of Corporate Activities”,
Journal of Accounting and Auditing Reviews, 40, 3-17.
[12] Min, S. H., Lee, J., and Han, I. (2006), “Hybrid genetic algorithms and
support vector machines for bankruptcy prediction”. Expert systems with
applications, 31: 652-660.
[13] Odom, M. D. and Sharda, R. (1990), “A Neural Network Model for
Bankruptcy Prediction”. IJCNN International Joint Conference on Neural
Networks. San Diego: CA, 2: 163-168.
[14] Reese, W. (1995), “Financial Analysis (second ed.)”. London: Prentice
Hall.
103
[15] Shah, J. and Murtaza, M. (2000), “A neural network based clustering
procedure for bankruptcy prediction”. American Business Review, 18 (2),
80-86.
[16] Varetto, F. (1998), “Genetic Algorithms application in the analysis of
insolvency risk”.
[17] Wallace Wanda, A. (2004), “Risk assessment by internal auditors using
past research on bankruptcy applying bankruptcy models”.
Vol. 2, No. 1, Ser. 5, (2014), 105-122
Estimation of Project Performance Using
Earned Value Management and Fuzzy
Regression
Mohammad Mahdi Asgari Dehabadi∗
Department of Industrial engineering,
University of Economic Science
Tehran, Iran
Mostafa Salari
Department of Industrial engineering,
Sharif University of Technology
Tehran, Iran
Ali Reza Mirzaei
Department of Technology Management,
Allameh Tabataba’ee University
Tehran, Iran
Abstract. Earned Value Management is a critical project management methodology that evaluates project performance from cost and
schedule viewpoints. The novel theoretical framework presented in this
paper estimates future performance of project regarding the past relative information. It benefits from fuzzy regression (FR) models in estimation process in order to deal with the vagueness and impreciseness of
real data. Furthermore, fuzzy-based estimation is evaluated using linguistic terms to interpret different possible condition of projects. The
proposed model can greatly assists project managers to assess prospective performance of project and alerts them in taking of necessary
actions. Finally, one illustrative case associated with a construction
project has been provided to illustrate the applicability of theoretical
model in real situations.
∗
105
106
M. M. Asgari Dehabadi, M. Salari, and A. R. Mirzaei
Keywords: Fuzzy regression; earned value management; project management; estimation.
1.
Introduction
Earned value management (EVM) is a project management technique
which reveals great capability in measurement of project performance
from different viewpoints. When EVM is properly applied, it provides
an early warning of performance problems. The PMBOK Guide initially defines EVM as “a management methodology for integrating scope,
schedule, and resources for objectively measuring project performance
and progress” [1]. The introduced indices of EVM make such measurement possible. Due to the simplicity and application of EVM systems in
different situations, many researchers applied the EVM in various organizations and projects [2-6]. On the other hand, there other researchers
discussed and improved the efficiency of EVM in real case projects [7-13].
Actually, there are two distinctive viewpoints for cost management
in an EVM system: Initially, it looks backward, measuring the past cost
and schedule performances of project via using cost performance index
(CPI) and schedule performance index (SPI), respectively. Secondly, it
looks forward proposing a process called estimate at completion process
(EAC and EACt) for estimation of project total cost and duration. Regarding the second viewpoint, EVM is a method for assisting project
managers to reach reasonable decisions concerning the future of ongoing projects. However, there are some situations in real case projects
that project managers require obtaining the cost future performance of
project in the upcoming milestones or to observe the future trend of cost
performance for taking necessary actions.
Hence, being aware of project total cost or duration is not enough for
taking managerial decisions. It seems that it would be an appropriate
idea to bridge the gap between these backward and forward viewpoints
of EVM which means to employ CPI and SPI for prediction of project future performances. However, there are many studies that just addressed
the estimation at completion process and attempted to improve their
obtained estimation. In this regard, [7, 10, 14]introduced planned value
Estimation of Project Performance Using ...
107
(PV), earned schedule (ES) and earned duration (ED) in order to develop distinctive models for prediction of project total cost. Moreover,
[15] utilized stochastic S-curves for forecasting of project performance. In
another study, [16] discussed cost estimation method in terms of effort
spent on a software project [17]introduced a final time and cost forecasting method applying statistical approach. [18] Developed a model
for estimation of project final cost concerning how to exceed the convergence to the appropriate result with less variation than typical model
for estimate at completion calculations. [19] Discussed a fuzzy neural
network to estimate at completion costs of construction project. [20]
Studied the accuracy of preliminary cost estimation in public work departments. Recently, [21] proposed a Bayesian approach to improve estimate at completion in earned value management.
To the best of authors’ knowledge, none of the researches in EVM
area of research attempted to take the advantage of CPI and SPI for
periodic estimation of project performance from cost and schedule view
point. Hence, the main contribution of this study is to concentrate on
this available lack in EVM technique and to develop a model which
is capable of project future performances. The rest of this study is
organized as follows:
Fuzzy theory is comprehensively described in section 2. It is followed
by introduction of EVM indices in section 3. Section 4 is dedicated to
explain how the fuzzy regression can be applied as a powerful tool to
predict EVM indices in future. The question related to the interpretation
of fuzzy-obtained values is responded in section 5. Eventually, in section
6, a case study is employed to show how the proposed can be utilized
for a real case project.
2.
Utilization of Fuzzy Theory in the Proposed
Model
In 1965, Lotfi Zadeh [22] introduced fuzzy set and theory to cope with
vagueness in systems where uncertainty increases due to fuzziness rather
than randomness. In doing so, the fuzzy theory utilizes different types
7.
119
Conclusion
A new method for prediction of project performance from two distinctive viewpoints, i.e. cost and schedule, is presented in this paper. Fuzzy
regressions as a powerful prediction tool and linguistic terms for interpretation of fuzzy values are then employed in the proposed model. Research
finding of applying the presented method for the case study indicates
that how efficiently the model is capable of providing initial warning in
the cases that the current condition of project performances are according to the plan but their general trend tend toward the weak sides. Using
integrated time series and simulation is recommended for further development of the proposed model.
Acknowledgments:
The authors would like to warmly thank Dr. Bagherpour for his helpful
comments and cooperation in this work.
References
[1] PMI. (2012), Practice standard for Earned Value management. PMI Publication.
[2] Al-Jibouri, S. H. (2003), Monitoring systems and their effectiveness for
project cost control in construction. International Journal of Project Management, 21 (2), 145-154.
[3] Bagherpour, M., Zareei, A., Noori, S., and Heydari, M. (2010), Designing
a control mechanism using earned value analysis: An application to production environment. International Journal of Advanced Manufacturing
Technology, 49 (5-8), 419-429.
[4] Baumeister, A. and Floren, A. (2011), Optimizing the Configuration of
Development Teams Using EVA: The Case of Ongoing Project Adjustments Facing Personnel Restrictions. International Journal of Information
Technology Project Management, 2 (1), 62-77.
[5] Moselhi, O., Li, J., and Alkass, S. (2004), Web-based integrated project
control system. Construction Management and Economics, 22 (1), 35-46.
120
[6] Oven, J. K. (2007), Implementing EV management in a R & D environment. AACE International Transactions EVM, 1, 01-05.
[7] Anbari, F. (2003), Earned Value Project Management Method and Extensions. Project Management Journal, 34 (41), 12-23.
[8] Cioffi, D. F. (2006), Designing project management: A scientific notation and improved formalism for EV calculations. International Journal
of Project Management , 24, 134-144.
[9] Jacob, D. S. (2003), Forecasting project schedule completion with earned
value metrics. The Measurable News, 1, 7-9.
[10] Lipke, W. (2003), Schedule is different. The Measurable News, 31-34.
[11] MoslemiNaeni, L. and Salehipour, A. (2011), Evaluating fuzzy earned
value indices and estimates by applying alpha cuts. Expert Systems with
Applications, 38 (7), 8193-8198.
[12] Salari, M., Bagherpour, M., and Kamyabniya, A. (2014), Fuzzy extended
earned value management: A novel perspective. Journal of Intelligent and
Fuzzy Systems. Pre-press.
[13] Salari, M., Bagherpour, M., and Wang, J. (2014), A novel earned value
management model using Z-number. International Journal of Applied Decision Sciences, 7 (1), 97-119.
[14] Jacob, D. S. and Kane, M. (2004), Forecasting schedule completion using
earned value metrics revisited. The Measurable News, 1, 11-17.
[15] Barraza, G. A., Back, W. E., and Mata, F. (2004), Probabilistic forecasting of project performance using stochastic S curves, 130 (1), 25.
[16] Dillibabu, R. and Krishnaiah, K. (2005), Cost estimation of a software
product using COCOMO II. 2000 model-A case study. International Journal of Project Management, 23 (4), 297-307.
[17] Lipke, W., Zwikael, O., Henderson, K., and Anbari, F. (2009), Prediction of project outcome: The application of statistical methods to earned
value management and earned schedule performance indexes. International Journal of Project Management, 27 (4), 400-407.
[18] Warburton-Roger, D. H. (2011), A time-dependent earned value model
for software projects. International Journal of Project Management, 29
(8), 1082-1090.
121
[19] Feylizadeh, M. R., Hendalianpour, A., and Bagherpour, M. (2012), A
fuzzy neural network to estimate at completion costs of construction
projects. International Journal of Industrial Engineering Computations,
3 (3), 477-484.
[20] Azman, M. A., Abdul-Samad, Z., and Ismail, S. (2013), The accuracy
of preliminary cost estimates in Public Works Department (PWD) of
Peninsular Malaysia. International Journal of Project Management, 31
(7), 994-1005.
[21] Caron, F., Ruggeri, F., and Merli, A. (2013), A Bayesian Approach to
Improve Estimate at Completion in Earned Value Management. Project
Management Journal, 44 (1), 3-16.
[22] Zadeh, L. (1965), Fuzzy sets. Information and Control, 338-353.
[23] Naeni, L., Shadrokh, S., and Salehipour, A. (2011), A fuzzy approach for
earned value management. International Journal of Project Management,
29, 764-772.
[24] Vandevoorde, S. and Vanhoucke, M. (2005), A comparison of different
project duration forecasting methods using EV metrics. International
Journal of Project Management, 24, 289-302.
[25] Tanaka, H., Uejima, S., and Asai, K. (1982), Linear regression analysis
with fuzzy model. IEEE Transaction on Systems, Man and Cybernetics,
12 (6), 903-907.
[26] Change, Y. and Ayyub, B. (2001), Fuzzy regression methods-a comparative assessment. Fuzzy sets and systems, 119 (2), 187-203.
[27] D’Urso, P. (2003), Linear regression analysis for fuzzy/crisp input and
fuzzy/crisp output data. Computational Statistics & Data Analysis, 42,
47-72.
[28] Yang, M. S. and Lin, T. S. (2002), Fuzzy least squares linear regression
analysis for fuzzy input-output data. Fuzzy Sets and Systems, 126, 389399.
[29] Kiossi, M. C. and Shapiro, A. (2006), Fuzzy formulation of the Lee-Carter
model for morality forecasting. Insurance, Mathematics and Economics,
39, 287-309.
122
[30] Sánchez, J. d. A. (2006), Calculating insurance claim reserves with fuzzy
regression. Fuzzy Sets and Systems, 157 (23), 3091-3108.

In the Name of God Shiraz Journal of System Management (SJSM

Transcription

Similar documents

Hexter-Fair FirstAmerican Title~

Appraisal Districts San Antonio Area

request to correct name or address on a real property account

Pope John Paul II`s Teaching on Sexuality and Marriage: An Appraisal

Preview

Tópicos de Contabilidade Gerencial

pharmacy - Cooper

$83,300.00

Estimating Home Value