A Simple Hybrid Particle Swarm Optimization

As a novel stochastic optimization technique, the Particle Swarm Optimization technique (PSO) has gained much attention towards several applications during the past decade for solving the global optimization problem or to set up a good approximate solution to the given problem with a high probability. PSO was first introduced by Eberhart and Kennedy [Kennedy and Eberhart, 1997]. It belongs to the category of Swarm Intelligence methods inspired by the metaphor of social interaction and communication such as bird flocking and fish schooling. It is also associated with wide categories of evolutionary algorithms through individual improvement along with population cooperation and competition. Since PSO was first introduced to optimize various continuous nonlinear functions, it has been successfully applied to a wide range of applications owing to the inherent simplicity of the concept, easy implementation and quick convergence [Trelea 2003]. PSO is initialized with a population of random solutions. Each individual is assigned with a randomized velocity based to its own and the companions flying experiences, and the individuals, called particles, are then flown through hyperspace. PSO leads to an effective combination of partial solutions in other particles and speedens the search procedure at an early stage in the generation. To apply PSO, several parameters including the population (N), cognition learning factor (cp), social learning factor (cg), inertia weight (w), and the number of iterations (T) or CPU time should be properly determined. Updating the velocity and positions are the most important parts of PSO as they play a vital role in exchanging information among particles. The details will be given in the following sections. The simple PSO often suffers from the problem of being trapped in local optima. So, in this this paper, the PSO is revised with a simple adaptive inertia weight factor, proposed to efficiently control the global search and convergence to the global best solution. Moreover, a local search method is incorporated into PSO to construct a hybrid PSO (HPSO), where the parallel population-based evolutionary searching ability of PSO and local searching behavior are reasonably combined. Simulation results and comparisons demonstrate the effectiveness and efficiency of the proposed HPSO. The paper is organized as follows. Section 2 describes the acronyms and notations. Section 3 outlines the proposed method in detail. In Section 4, the methodology of the proposed HPSO is discussed. Numerical simulations and comparisons are provided in Section 5. Finally, Concluding remarks and directions for future work are given in in Section 6. O pe n A cc es s D at ab as e w w w .ite ch on lin e. co m