Making efficient AI models for games with imperfect information can be a particular challenge. Considering the large number of possible moves and the incorporated uncertainties building game trees for these games becomes very difficult due to the exponential growth of the number of nodes at each level. This effort is focused on presenting a method of combined Case Based Reasoning (CBR) with AI Planning which drastically reduces the size of game trees. Instead of looking at all possible combinations we can focus only on the moves that lead us to specific strategies in effect discarding meaningless moves. These strategies are selected by finding similarities to cases in the CBR database. The strategies are formed by a set of desired goals. The AI planning is responsible for creating a plan to reach these goals. The plan is basically a set of moves that brings the player to this goal. By following these steps and not regarding the vast number of other possible moves the model develops Game Trees which grows slower so they can be built with more feature moves restricted by the same amount of memory.
|Title of host publication||Proceedings of the 3rd International Conference on Digital Interactive Media in Entertainment and Arts (DIMEA'08), 10-12 September 2010, Athens, Greece|
|Place of Publication||New York|
|Publisher||Association for Computing Machinery, Inc|
|Publication status||Published - 2008|