Title
Simulation and Analysis of Strategy in a Variation of the Gobblet Game
Faculty Mentor(s)
John Holliday and Dianna Spence
Location
Library Technology Center David L. Potter Special Collections Room 382
Start Date
27-3-2012 11:00 AM
End Date
27-3-2012 12:15 PM
Description/Abstract
The purpose of our research is to apply graph theory and computer simulation using CUDA to explore the strategy of the game Gobblet Gobblers. This game is simpler than Gobblet but more sophisticated than tic tac toe. On a 3 x 3 grid, each player has three sized pieces, and larger pieces may be used to cover either player’s smaller pieces. In addition, pieces can be moved to another position on the board once played. A bipartite graph models the mapping between winning combinations and positions on the board. This graph model is used with a combinatorial approach to examine offensive and defensive strategies. The computer simulation is intended to produce exhaustive outcomes given a specific strategy. CUDA was chosen because of its ability to execute many scenarios in parallel. Tentative results include optimal beginning strategies for player one, as well as moves that put either player at a disadvantage. Faculty Advisers: John Holliday and Dianna Spence.
Simulation and Analysis of Strategy in a Variation of the Gobblet Game
Library Technology Center David L. Potter Special Collections Room 382
The purpose of our research is to apply graph theory and computer simulation using CUDA to explore the strategy of the game Gobblet Gobblers. This game is simpler than Gobblet but more sophisticated than tic tac toe. On a 3 x 3 grid, each player has three sized pieces, and larger pieces may be used to cover either player’s smaller pieces. In addition, pieces can be moved to another position on the board once played. A bipartite graph models the mapping between winning combinations and positions on the board. This graph model is used with a combinatorial approach to examine offensive and defensive strategies. The computer simulation is intended to produce exhaustive outcomes given a specific strategy. CUDA was chosen because of its ability to execute many scenarios in parallel. Tentative results include optimal beginning strategies for player one, as well as moves that put either player at a disadvantage. Faculty Advisers: John Holliday and Dianna Spence.