A Novel Adaptive PSO Algorithm on Schaffer's F6 Function
Xiaohong Qiu, Jun Liu · 2009
Analyzing the distance between the location and the new location, we conclude inertia weight method which linearly decreases from 0.9 to 0.4 has the powerful local search ability on Schafferpsilas F6 function. In order to improve the balance between local and global search ability, the novel adaptive PSO algorithm which evaluates a reset function to control the inertia weight value is proposed. Once plunged into the local optimum, inertia weight, pbest and gbest should be reset to get away from the local optimum. Compared with the particlepsilas traces, the novel algorithm has a great potential advantage. Simulation results show that the novel adaptive algorithm is better than the inertia weight algorithm in terms of the successful searching rate on Schafferpsilas F6 function.