Probabilistic stochastic diffusion search

Abstract

Stochastic Diffusion Search (SDS) is a population-based, naturally inspired search and optimization algorithm. It belongs to a family of swarm intelligence (SI) methods. SDS is based on direct (one-to-one) communication between agents. SDS has been successfully applied to a wide range of optimization problems. In this paper we consider the SDS method in the context of unconstrained continuous optimization. The proposed approach uses concepts from probabilistic algorithms to enhance the performance of SDS. Hence, it is named the Probabilistic SDS (PSDS). PSDS is tested on 16 benchmark functions and is compared with two methods (a probabilistic method and a SI method). The results show that PSDS is a promising optimization method that deserves further investigation. © 2012 Springer-Verlag.