Chomp The Graph!
Transcription
Chomp The Graph!
By: Aylin Perez January 12, 2014 Table of Contents I. Research Question 1 II. Introduction 1 III. History 2 IV. Mathematical Background 3 V. Investigation 7 VI. Conclusion 15 VII. Applications and Extensions 18 VIII. References 19 -2- I. Research Question: Given a finite undirected graph G, two players alternate turns and remove either a single edge E or a vertex V with all incident edges. Whoever removes the last vertex, leaving their opponent with the empty graph, wins. Which player has the winning strategy for games on trees, cycles, and complete graphs? II. Introduction: The game Chomp the Graph is an excellent, fun way to introduce graph theory to beginners. Why? It may seem simple enough, but it turns out to be a game where one has to think critically and strategically to win. Every move counts. The underlying mystery and my recent study of graph theory is what drove me to do more research on the game. Which player will always, sometimes, or never win? Why? I hypothesized what the answer to the question would be prior to the investigation in order to compare my hypothesis with my findings at the end. This provided me with a picture as to what I was expecting and lead to more excitement, an enthusiastic approach, and surprise when I finally arrived at my answer. Before I started investigating, I believed that the first player would always win and have the winning strategy. Since the first player would lead the game, he or she can use that to his or her advantage. My investigation, however, would prove me wrong! -1- III. History: Leonhard Paul Euler was a Swiss mathematician and one of the most prolific mathematicians of all time, despite suffering from blindness later on in his life. He is credited for being the first mathematician to use graph theory around 1736 to solve the infamous Königsberg Bridge problem. The problem took place in the city of Königsberg, which had seven bridges connecting two islands with the main land. The problem questioned whether someone could walk across all seven bridges once and only once. Euler was the first mathematician to arrive at a solution in terms of graph theory and proved that no matter where a person started, it was impossible to walk across the bridges once and only once. In order to do this, he exchanged landmasses for vertices and represented the bridges as edges connecting the vertices. Euler not only simplified the problem, but also explained why it wasn’t possible. If one were to take a pencil and attempt to trace over all the edges, it would be impossible to do so without repeating any or picking up one’s pencil. To show why, Euler came up with the concept of degrees, in which he assigned each vertex a number based on the edges extending from that vertex. Euler proposed that the only way each edge could be traversed once and only once for any graph would be if at most 2 vertices had odd degrees. Any other combination would not yield what we now call Eulerian Graphs. -2- Chomp the Graph is a finite game. In other words, the game cannot go on forever: one player will either win or lose. The game is very similar to the popular game Nim, which is an impartial game where two players take turns taking stones from different heaps and can take as many they want as long as the stones are from the same heap. The object of the game is to leave the opponent with nothing to take. Similarly, Chomp the Graph is a game where one wins when the opposing player has nothing else to take. However, instead of only stones and heaps, vertices and edges are taken into account. IV. Mathematical Background Before starting the investigation, it is important to know all the mathematical concepts that are involved in the game. This serves as a way to get a deeper understanding behind the strategy and why either player possesses it. Graph Theory- Basics A graph G consists of vertices V and edges E. The number of vertices is referred to as order, while the number of edges is referred to as size. Each vertex has a number of degrees, which is the number of edges that extend from that vertex. The example to the right is a graph with order 5, as well as size 5 since there are 5 edges in total. -3- Isomorphism Two graphs can be isomorphic, which means that they are essentially the same graph but may look different. In order to determine isomorphism, order for both graphs must be the same as well as size, and most importantly, there must be a one-to-one correspondence between vertices. For instance, the two graphs below are isomorphic because not only do they have the same order and size, but there is a one-to-one correspondence between the vertices. Isomorphism is important because when listing all the possible combinations of how the game Chomp the Graph could turn out, there will be graphs that will be isomorphic and only one should be counted. Directed Graphs Directed graphs are graphs where the edges have direction. These graphs are not part of the investigation nor the game, but it is important to know them. Cycles & Complete Graphs A cycle graph is a circuit without repeated vertices except at the ends. Furthermore, it is a graph containing a single cycle through all vertices. These graphs are commonly labeled as Cn where n is the number of vertices. They are also two-regular, meaning each vertex has a degree of 2. When examining cycle graphs in the investigation, they will be undirected, such as C3 and greater. -4- Graphs are complete when each pair of vertices is connected by one edge. Furthermore, every vertex is connected to all other vertices but itself. If that’s the case, then the number of edges of each vertex will be one less than the number of vertices. Therefore, to figure out the total number of edges, one would multiply the order, n, with one less than the number of vertices, and divided it by two to avoid counting twice. This process yields the formula: This formula is very important when determining the number of edges for complete graphs and will be useful during the investigation. Complete graphs are commonly referred to as Kn, where n is the number of vertices. Trees and Bipartite Graphs Trees are graphs with no cycles where any two vertices are connected by exactly one path. Trees have many applications, and are very useful in chemistry, when drawing hydrocarbons. A tree with n vertices always has n-1 edges, a fact that will help in the investigation. This statement can be proven by using induction: i. Base Case: a. A tree of order 1 has no edges, which is one less than the number of vertices. ii. Inductive Step: a. Assume a tree graph with n vertices has n-1 edges for all n=k b. Prove that a tree with k+1 vertices has k edges. Let Tk+1 be any tree with k+1 vertices. Consider Tk, the subgraph created by removing a vertex v of degree one and the edge connecting it to the graph. By using “connected subgraph of -5- a tree is a tree”, Tk must also be a tree. Since it was assumed that a tree graph with n vertices has n-1 edges for all n=k, then Tk must have k-1 edges. That results in Tk+1 having k edges. Therefore, a tree with n vertices has n-1 edges for all n. Bipartite graphs are graphs whose vertices can be divided into two disjoint sets such that no two graph vertices are adjacent. An easy way to determine if a graph is bipartite is by choosing two colors and coloring vertices in a way such that only vertices of opposing colors are connected. Since all trees are bipartite, the winning strategy for trees will apply to certain bipartite graphs. In order to determine if the bipartite graph is a tree or not, there has to be at least one cycle in the graph for it not to be a tree. The Winning Strategy The winning strategy for Chomp the Graph is effective throughout the game, as long as all of the rules are met. The key is to count the number of edges and vertices. The player wants to take one of whichever category has an odd number; in other words, if there are 4 vertices and 3 edges in total, he or she should remove an edge. Furthermore, when removing a vertex, one should remove the vertex with the highest degree. Last but not least, if both order and size are odd, one should go with the highest odd number. For instance, if the game is at 3 vertices and 1 edge, the vertex with the highest degree should be removed. After all of this information, it’s time to investigate! -6- V. Investigation: Tree Graphs: Order 2 Player 1- 1 win Player 2- 1 win Order 3 Player 1- 4 wins Player 2- 3 wins -7- Order 4 Player 1- 20 wins Player 2- 19 wins Order 5 Player 1- 53 wins Player 2- 52 wins -8- # of Vertices (Order) Total Combinations Player 1 wins Player 2 wins 2 2 1 – 50% 1 – 50% 3 7 4 – 57.1% 3 – 42.9% 4 39 20 – 51.3% 19 – 48.7% 5 105 53 – 50.5% 52 – 49.5% Based on the results, player 1 has the winning strategy. As shown in the diagram above, player 1 always has a slightly higher percentage in winning out of all the possible ways the game could turn out. Player 2 can’t win unless player 1 messes up. The definition of trees serves as the reason of why player 1 has the winning strategy. As proven before, a tree with n vertices has n-1 edges. Thus, player 1 has the advantage of starting with an even-odd pair every time, despite what value n is. Thus, if player 1 follows the strategy, he or she will always win. Cycle Graphs (Note that C2 is directed and thus not included.) Order 3 Player 1- 4 wins Player 2- 5 wins -9- Order 4 Player 1- 22 wins Player 2- 24 wins Note that this graph can also serve as an example of a bipartite graph that is also a tree. Player 2 is the dominant player. Order 5 Player 1- 156 Player 2- 161 - 10 - # of Vertices (Order) Total Combinations Player 1 wins Player 2 wins 3 9 4 – 44.4 % 5 – 55.6% 4 46 22 – 47.8% 24 – 52.7% 5 317 156 – 49.2 % 161 – 50.8% Cycles prove to be the opposite of trees; player 2 has the winning strategy! As shown in the diagram above, player 2 is the dominant player, as he or she wins the most games out of all the possible combinations in the different cases. The design of cycle graphs is the reason why player 2 wins all the time, despite the number of vertices. As seen in the diagrams above, the game always starts with an equal order and size, leaving player 1 no choice; he or she can’t use the strategy. This leaves player 2 with an odd-even pair, and the capability of using the winning strategy. Complete Graphs Note that K2 and K3 are T2 and C3 respectively. Order 4 Player 1- 90 Player 2- 96 - 11 - # of Vertices (Order) Total Combinations Player 1 wins Player 2 wins 2 2 1 – 50% 1 – 50% 3 9 4 – 44.4 % 5 – 55.6% 4 186 90 – 48.4% 96 – 51.6% K3 is a graph where player 2 has the winning strategy, as proved above with cycles. Since mapping out all of the possible ways Chomp the Graph could turn out on a complete graph with order 5 and 6 yield an incredible amount of combinations, let’s think about the game from a strategic point of view: Order 5 - 12 - Even though the first player has the odd-even pair at the beginning of the game on a K5 graph, we see that player 2 could also potentially win! Complete graphs are the only graphs where the player who doesn’t have the strategy can counterattack! Since K3 is a graph where player 2 has the winning strategy, he or she wants to put player 1 in the position of making a move on a K3 graph. As shown in the diagram above, whenever that occurs, player 2 wins. K4 is another graph where the data shows the possibility of player 2 winning is high, but player 1 can counterattack! Player 1 starts with an even-even pair of vertices and edges, 4 and 6 respectively. Even though it might seem like player 1 is stuck, he or she can always remove one vertex to form a K3 graph and thus, the second player can’t use the strategy; whichever move he or she makes will always result in player 1 winning. Therefore, to add to the winning strategy for complete graphs, I’ve discovered that the goal of one player is to put the opposing player in a position where he or she is stuck with a K 3 graph. That will guarantee that the original player wins every time. To demonstrate this, let’s look at a K6 graph. As seen in the following page, player 1 wins more times that player 2, and it’s all about forcing K3 on the opposite player and following the other rules that have been established. Based on the counterattack, which cannot be used in any other graph but a complete graph, I conclude that complete graphs are the fairest graphs to play on when the number of vertices is greater than or equal to 5. Not only is the winning strategy useful, but the counterattack proves to be even better. As seen in the following graph, just the winning strategy isn’t always helpful; the counterattack, however, guarantees a win. - 13 - - 14 - VI. Conclusion To conclude, Chomp the Graph is a game that is won by using the following strategy: Count the number of vertices and edges; remove one of the categories that have an odd number. If both categories are odd-numbered, the player should remove the one with the largest odd number. If both categories are even-numbered, the player whose turn it is will lose; in other words, the player with either the odd-odd (different) or odd-even pair will win. If both categories are odd-numbered and the numbers are the same, the player whose turn it is will lose; in other words, the next player, who will always be left with the odd-even pair, will win. In order to determine which player has the winning strategy, all the combinations of how the game could turn out for different types of graphs were mapped, without loss of generality. The percentages signaled which player had the winning strategy, where the player with the highest percentage, even by a slight amount, was able to use the strategy. Meanwhile, the other player had no way of counterattacking unless the original player messed up. This applied to trees and cycles, while complete graphs yielded a different result. - 15 - Trees: Player 1 will always win using the strategy. As shown below, Player 1 always has a slight advantage. # of Vertices (Order) Total Combinations Player 1 2 2 1 – 50% 3 7 4 – 57.1% 3 – 42.9% 4 39 20 – 51.3% 19 – 48.7% 5 105 53 – 50.5% 52 – 49.5% Player 2 1 – 50% Cycle graphs: Player 2 has the winning strategy as shown by the percentages below. # of Vertices (Order) Total Combinations Player 1 Player 2 3 8 3 – 37.5% 5 – 62.5% 4 46 22 – 47.8% 24 – 52.2% 5 317 156 – 49.2 % 161 – 50.8% Complete Graphs: Complete graphs proved to be an exception. In addition to the strategy mentioned above, there was a form of counterattack on complete graphs. The player wants to force a K3 situation on the opposing player to guarantee a win. Based on this, it was concluded that Chomp the Graph is the fairest to play on a complete graph with order ≥ 5 since the strategy mentioned above is not enough. - 16 - # of Vertices (Order) Total Combinations Player 1 Player 2 2 2 1 – 50% 1 – 50% 3 8 3 – 37.5% 5 – 62.5% 4 186 90 – 48.4% 96 – 51.6% The following chart sums up the results: Guaranteed Win Graphs Trees Player 1 Player 2 (any order) Cycles (any order) (only when order=2 and 4 with the Complete (only when order=3) counterattack) - 17 - VII. Applications and Extensions Chomp the Graph is a great way to introduce Graph Theory to students who are beginning to study the topic. It can serve as a great lesson, where students will learn to think outside of the box and think in terms of vertices and edges. Chomp the Graph doesn’t stop here! o The original game can be changed by giving each player the choice of one or two moves. For example, if one player decides to play only once in one round, the next time they go, they have to play twice. However, this resets, so the next turn, that player can go twice again, but then would have to play only once the next time. This will make the problem more interesting because it will change the approach of either player toward winning the game. Now, they would have to be conscious of the move they make and the move they have left to make. Who has the winning strategy and in what graphs? o Another variation of the game would be varying the number of vertices/edges each player can take per turn. For instance, if one player takes a vertex, when he plays again, he has to take an edge. It can switch every turn or there could be a limited number of vertices/edges. Which player will win and in what graphs? o One final variation is using directed graphs! This will re-define the game, because removing a vertex will not remove all the edges attached to it; there would be a choice: the edges going facing away from the vertex or the edges facing the vertex. What is the winning strategy, and which player has it on what graphs? - 18 - VIII. References A. Bogomolny, Mathematicians Doze Off from Interactive Mathematics Miscellany and Puzzles. http://www.cut-the-knot.org/do_you_know/DozingOff.shtml Draisma, Jan, and Sander Van Rijnswou. How to Chomp Forests, and Some Other Graphs (n.d.): n. pag. Mathematisches Institut. Web. “Nim." Encyclopedia Britannica. Encyclopedia Britannica Inc., n.d. Web. <http://www.britannica.com/EBchecked/topic/415458/nim>. Rhishikesh. Graph Theory Origin and Seven Bridges of Königsberg (n.d.): n. pag. Georgia State University. Web. “The Origins of Graph Theory." (n.d.): n. pag. University of Kansas. Web. <https://www.math.ku.edu/~jmartin/courses/math410-S09/graphs.pdf>. Weisstein, Eric W. "Complete Graph." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/CompleteGraph.html Weisstein, Eric W. "Cycle Graph." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/CycleGraph.html Weisstein, Eric W. "Tree." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/Tree.html - 19 -