An analysis of PSO inertia weight effect on swarm robot source searching efficiency

Abstract

Particle swarm optimization (PSO) is a well-known metaheuristic and population based optimization algorithm. PSO has been used in many science and engineering applications. One of the examples include as a source searching algorithm in swarm robotics. In general, PSO searching performance is determined by its parameters include inertia weight and acceleration coefficients. In this paper, a comparison analysis of inertia weight adjustment strategies on the source searching capability is presented. Several methods of adjusting the inertia weight are studied such as constant inertia weight, randomized inertia weight, linearly decreasing inertia weight, increasing inertia weight, and adaptive inertia weight. The analysis is performed by keeping other PSO parameters including the acceleration coefficients constant. The comparison is benchmarked based on a few performance indexes:convergence time, accuracy, percentage of search success and smoothness of trajectory. The comparison results presented in this study are useful as a basic guideline for development of a better PSO based source searching algorithm for swarm robotics in the future.