A PSO Optimization Scale-Transformation Stochastic-Resonance Algorithm With Stability Mutation Operator

When using the PSO (particle swarm optimization) optimization adaptive stochastic-resonance method, the initial value and value range of the optimization parameters are defined inappropriately, divergence problems may easily emerge in the calculation process, and optimization may stop prematurely. To solve this problem, this research has analyzed the parameters that influence system stability using the scale-transformation stochastic-resonance solution procedure, and the value range leading to algorithm stability was obtained. On this basis, a stable mutation operator has been proposed, which is used in mutation operations on particles outside the stable condition. To ameliorate the poor local search ability and low convergence speed of the PSO algorithm in the later iteration stage, an inertial weight degression strategy based on a particle distance index has been developed. Based on these two research results, a PSO optimization scale-transformation stochastic-resonance algorithm with mutation operator has been proposed. The proposed algorithm has been used to detect numerically simulated signals and rotor test-table data. The results show that when the stable mutation operator acts on the SR optimization parameters, divergence is effectively avoided, and the stability of the iterative algorithm is improved accordingly. By adding the inertial weight degression strategy to the PSO algorithm, iteration speed could be improved at the same time.