A Competitive Social Spider Optimization with Learning Strategy for PID Controller Optimization

Tuning the parameters of PID controller is a difficult problem, since it is hard to get the optimum parameters by the traditional methods, new methods are required. Nature-inspired algorithms perform powerfully and efficiently on global optimization problems. Social spider optimization (SSO) is one of the novel nature-inspired algorithms, and it exhibits good performance on avoiding premature convergence. However, the efficiency of SSO degrades when used in applications such as PID controller optimization whose objective function with highly correlated variables. In order to overcome this disadvantage, based on the SSO, a competitive social spider optimization (CSSO) is proposed in this paper. To enhance the performance of SSO, we regroup the spiders and the diversity of population is increased. Inspired by the competitive mating behavior of spiders, the competitive mating mechanism is introduced, and a learning strategy is used for the new born spider. The CSSO is applied to optimize the parameters of PID controller, and the simulation results show that the performance of CSSO is promising in PID controller optimization.