Hybrid particle swarm optimization incorporating fuzzy reasoning and weighted particle

In conventional particle swarm optimization (PSO), the search behavior has two principal forces of the moving direction to guide the particles toward their personal best (pbest) and the global best (gbest) positions. However, if the particle lies too close to either pbest or gbest, the optimization of the swarm is likely to be trapped into a local optimum. To overcome the local optimum problem, this paper proposes a hybrid particle swarm optimization incorporating fuzzy reasoning and a weighted particle (HPSOFW) to establish a novel search behavior model to improve the searching capability of the conventional PSO algorithm. In the proposed search behavior model, a weighted particle is incorporated into the algorithm to modify the searching direction and fuzzy reasoning is used to adjust an attraction factor and inertia weight such that the particle has a better opportunity to find the optimal solution. Based on adjustment of the attraction factor and inertia weight, the proposed search behavior model takes into consideration of both search strategies of exploitation (local search) and exploration (global search) during the optimization. Simulation results show that the proposed HPSOFW has much better performance than that of the existing optimization algorithms for ten benchmark functions. To demonstrate its feasibility, the proposed HPSOFW is also applied to the learning of neural network for nonlinear system modeling before applying it to model an energy consumption system with satisfactory performance.