An adaptive particle swarm optimization algorithm with newton-steepest descent operator

In order to solve the problem that particle swarm optimization is easy to fall into local optimal solution, low convergence accuracy and convergence-rate slowly in the later stage, the Newton-steepest descent operator, dynamic inertia weight and influence degree decision are introduced into the updating of particle swarm optimization. An adaptive particle swarm optimization algorithm combines with Newton-steepest descent operator (NSWPSO) is proposed. The improved algorithm, standard particle swarm optimization algorithm, adaptive inertia weight particle swarm optimization algorithm and linear decreasing inertia weight particle swarm optimization algorithm are applied to 12 test functions of different dimensions at the same time, by the comparing and analyzing search results, the analysis of T-test difference and the analysis of the optimization rate with average iteration times, where the 10-dimensional test function reaches the expected value. It can make the improved-algorithm stable and fast to find the global optimum-solution.