Multi-objective Optimization and Computational Intelligence

Abstract

Computational intelligence is a comprehensive name embracing not only artificial intelligence but also several recent topics such as artificial neural networks, evolutionary methods, metaheuristics using swarm intelligence (e.g., particle swarm optimization, ant colony optxximization), fuzzy logic and so on. Computational intelligence techniques are now widely applied in various practical fields. In this book also, we use several techniques of computational intelligence for sequential multiobjective optimization. This chapter will be concerned with machine learning which is utilized commonly for knowledge acquisition. Typical approaches to machine learning are (1) to find an explicit rule as if–then rule and (2) to judge newly observed data by an implicit rule which is usually represented as a nonlinear function. Well-known ID3 (recently C5.0) [115] and CART [14] belong to the former category. On the other hand, artificial neural networks and mathematical programming approaches belong to the latter category. In this book, we focus on the latter category