An improved adaptive particle swarm optimization approach for multi-modal function optimization

This paper introduces an improved adaptive particle swarm optimization (APSO) technique for locating the global minima of multi-modal functions. The APSO extends the original PSO with an improved search ability which is achieved by adaptively adjusting the inertia weight of each particle with respect to the objective function values. Further we have introduced the concept of re-initialization of a part of the population which makes the algorithm to converge to the global optimum. We test our APSO for several multi-modal functions to find the global optima. The proposed APSO can successfully locate the global optima of all the test functions.