Genetic-based fuzzy adaptation

A novel paradigm for adaptation in discrete event dynamic systems control is presented. In this approach a few optimal control policies are used to train an adaptation module capable of generalizing over several operating conditions. The general idea is then applied, through simulations, to an inventory system. A fuzzy adaptation module is built from a few optimal examples generated using a genetic algorithm. The fuzzy module is then capable of adapting the inventory policy achieving much better results than a static policy