An IB-PSO algorithm for unconstrained nonlinear programming problems

Particle swarm optimizer (PSO) is an efficient algorithm to find the best solution in engineering. However, the classical particle swarm optimization algorithm and the existing variational algorithm still have the problems of prematurity and slow convergence speed in the late iteration in different application problems. To make up the above demerits, we proposed a hybrid algorithm PSO to combine an improved PSO (IPSO) with an improved Broyden-Fletcher-Goldfarb-Shanno (BFGS) method. The hybrid algorithm introduced the convergence direction of the RL-BFGS to correct the evolution direction and enhance global search capability. And the initial point of the RL-BFGS is decided by the IPSO. For the sake of the effectiveness evaluation of IB-PSO, The performance of eight improved PSO algorithms was compared by optimizing six benchmark functions with different characteristics, such as single mode, multi-mode and rotation. The results show that compared with the other eight algorithms, the proposed IB-PSO has good performance in solving multimodal problems. Furthermore, the proposed approach can be used in engineering problems to obtain high-quality solutions.