A New Optimization Technique

This paper describes a new optimization technique to locate the minimum or maximum value of a function. The new technique, which we call the particle swirl algorithm (PSA), utilizes multiple solutions ("particles") that rotate or "swirl" around a center point. The swirling motion is based on the principle of water spiraling toward a drain, so the center point is known as the vortex. The vortex is not stationary, but rather continually moves towards the highest known fitness value in the search space. Since the particles are drawn towards the vortex, the search is concentrated in areas that have previously yielded good results. We tested the particle swirl algorithm on three different functions with ten different peak (target) locations and then compared the results against a gradient-based optimization technique and the particle swarm optimization technique. In all cases (all three functions and all peak locations), the PSA performed equal to or better than the other optimization techniques