By merging the theory of moves TOM and the fuzzy sets theory, we have developed the ลฝ . theory of fuzzy moves TFM to make better fuzzy moves for fuzzy games. Since the data granularity of conventionally used fuzzy sets is too low to contain more heuristic information and mined knowledge, we take primary fuzzy sets with higher data granularity as fundamental elements for fuzzy reasoning so as to make more reasonable moves.
ลฝ . The simulation results indicate that 1 TFM with normal fuzzy reasoning can make better and more reasonable moves than TOM with precise reasoning since different ลฝ . global strategies are taken into account by TFM and 2 the novel fuzzy reasoning methodology is more reasonable and more useful to make fuzzy moves than the conventional one. แฎ 1999 John Wiley & Sons, Inc.
4 แ 13 ods of the theory of moves TOM and fuzzy set theory. Some methods of making fuzzy moves were based on the conventional fuzzy reasoning methodology using Zadeh's fuzzy sets. 1,2,12,13 For example, a conventional fuzzy reasoning method was used to perform a nonlinear transformation mapping from a local game to a global game, 2 and neural networks were applied to learn such a nonlinear mapping from sample data. 1 Now, we use a more reasonable fuzzy reasoning methodology to perform nonlinear game transformations.