Performance of the Swarm-based Multiobjective Optimization Algorithms under Chaotic Noisy Problems

Abstract

– Chaotic systems are non-linear dynamic real-life systems which has randomized nature that cannot be modelled, and chaotic maps are functions that generate a chaotic behavior from a relatively simple formulation. Chaos can be observed at the real-life engineering systems and generally the chaotic behavior of these systems omitted due to the insufficient mathematical tools and irregular nature of the chaotic influence. Since they are existing and can be considered in the engineering system. Chaotic maps can be used to generate random numbers. Because of the chaotic nature of this randomize data it is hard - impossible- to handle these signals. The chaotic maps can be used as noise, and in this research, it is applied to the objective functions to generate chaotic noise, and the problems set is named as chaotic noisy benchmark problems (CNBP). In this research the performance of the swam-based multiobjective optimization algorithms is evaluated and analysis under CNBP. The solution for the question “Can evolutionary algorithms solve CNBP?” will be answered. It is showed after the empirical studies that the Chaotic map-oriented random numbers are relatively hard to handle when compared with Gaussian noise.