Evolutionary multiobjective optimization and multiobjective fuzzy system design

Abstract

Evolutionary multiobjective optimization (EMO) is one of the most active research areas in evolutionary computation. EMO algorithms have been successfully used in various application areas. Among them are multiobjective design of neural networks and fuzzy systems. Especially, fuzzy system design has often been discussed as multiobjective problems. This is because we have two conflicting objectives in the design of fuzzy systems: accuracy maximization and complexity minimization. In this paper, we first explain some basic concepts in multiobjective optimization, a basic framework of EMO algorithms and some hot research issues in the EMO community. Next we explain EMO-based approaches to the design of fuzzy systems. We demonstrate through computational experiments that a large number of non-dominated fuzzy systems with different accuracy-complexity tradeoffs can be obtained by a single run of an EMO algorithm. Then we describe the use of EMO algorithms in other areas such as neural networks, genetic programming, clustering, feature selection, and data mining.