LARES: An Artificial Chemical Process Approach for Optimization

This article introduces a new global optimization procedure called LARES. LARES is based on the concept of an artificial chemical process (ACP), a new paradigm which is described in this article. The algorithm's performance was studied using a test bed with a wide spectrum of problems including random multi-modal random problem generators, random LSAT problem generators with various degrees of epistasis, and a test bed of real-valued functions with different degrees of multi-modality, discontinuity and flatness. In all cases studied, LARES performed very well in terms of robustness and efficiency.