DESIGNING FDI OBSERVERS BY IMPROVED EVOLUTIONARY MULTI-OBJECTIVE OPTIMIZATION

The paper considers using evolutionary multi-objective optimization (EMO) for solving FDI design problems. In order to prevent genetic procedures from premature convergence and to increase the effectiveness of searching for sought optimal solutions, different types of niching allow taking care of their diversity in consecutive generations. Another problem results from considering design objectives of high dimensions, when the conception of Pareto-domination is not effective. The paper discusses the characteristics of several innovative mechanisms concerning the methods of selection and parental crossover, based on concepts of niches and genders.