On the performance of standard nature-inspired algorithms in solving non-stationary optimization problems

Abstract

Non-stationary optimization problems are a very important class of problems in many practical applications. These problems are characterized by objective functions and constraints that change with time or environmental conditions, so the optimization solution also needs to be dynamically adjusted accordingly. Many algorithms from the field of evolutionary and biology inspired computation are known as an effective approach for dealing with hard optimization problems in changing environments, that is the result of modelling of self-organized systems in nature and evolution in the biology. Natural systems always exist in the changing environments. This article aims to compare the performance of three common nature-inspired techniques, namely genetic algorithm (GA), particle swarm optimization (PSO) algorithm, and differential evolution (DE) in their standard implementation when solving non-stationary optimization problems, so as to provide a reference and rationale for subsequent selection of appropriate algorithms and improvements.