Fuzzy reasoning and genetic algorithms for decision making problems in uncertain environment

In the present paper, Genetic Algorithms (GA) have been used to optimize the design of a generic expert system based on fuzzy logics. Using the classical branch-and-bound method, the latter performs a heuristic graph search to solve a decision-making problem under uncertain environment. After building this fuzzy expert system, we isolate a set of parameters which are important for the system efficiency, and we show how these parameters can be optimized `automatically' using Genetic Algorithms. This optimization brings significant improvement over the manual tuning of the parameters in the specific case of a fuzzy expert system developed for the automatic target recognition of armoured vehicles starting from short-range infra-red images.