Managing of cooperative genetic algorithms by intelligent agent

Genetic algorithms (GAs) are very popular optimization tool, although efficient applications of GAs requires users have problem with setting their parameters and used genetic operators to obtain satisfactory solution in acceptable time. We propose an intelligent agent as a control mechanism for a group of cooperating genetic algorithms. A core of the system is a family of cooperating genetic algorithms. Manager, a kind of fuzzy expert system, is responsible for control of GAs to assure effectiveness of the search process. It requires providing knowledge concerning influence of some parameters of GAs on tempo and mode of evolution. Such knowledge in the form of fuzzy inference rules should be included into Manager. Analyser - the third part of the system, will gather knowledge concerning progress and actual state of a working system. This knowledge will be used in the form of facts to fire suitable rules. The simulation study of efficiency of developed system is presented and discussed.