AN EFFICIENT HYBRID DIRECT VAGUE FUZZY MOVES SYSTEM USING FUZZY-RULES-BASED PRECISE RULES

In this paper a game transformation is proposed to enable a player to get a globally reasonable goal which is not only as advantageous to hi s own side as possible but also as disadvantageous to his opponent as possible. The Theory of Fuzzy Moves (TFM), based on the Theory of Move s (TOM) and game theory, is suitable for making a player reach a relat ively good fuzzy outcome. Through analysis, we have found that only so me of the fuzzy moves' rules are vague (i.e. fuzzy), and the others ar e direct (i.e. precise). We have found that (1) for the direct rules, the hybrid system is fault-tolerant and stable; (2) for the vague rule s, the hybrid system is sensitive for the relevant players' fuzzy payo ffs and globally strategic goals since there exist sensitive points or curves. Through the mathematical simplification of the vague rules' i nference, we can finally construct the hybrid direct-vague fuzzy moves system by directly using the simple mathematical mapping functions wi thout using the conventional fuzzification, fuzzy inference, and defuz zification. Importantly, the hybrid direct-vague fuzzy moves system wi th the simple mapping functions can run much more quickly than the fuz zy moves system with the complex fuzzy mapping functions. Finally, som e typical examples of globally optimal fuzzy moves have shown that the hybrid direct-vague fuzzy moves system is stable and fault-tolerant e nough to deal with the fuzzy moves in dynamic and uncertain situations . (C) 1997 Elsevier Science Ltd. All rights reserved.