진화 알고리듬을 이용한 항공기 제어기 설계

Abstract

An aircraft acceleration control system is designed using various evolutionary algorithms. Evolution strategies are used to find the optimal gains for the controller. We also apply co-evolutionary algorithm in finding the saddle point solutions for the same system. Worst-case evolutionary algorithm is proposed for designing a robust controller. In the proposed algorithm, one group co-evolves while the other group simulates a random behavior. The evolving group is able to find a robust value for the search space. The solutions and properties of three methods are compared and analyzed.