Quantum Particle Swarm Optimization Based Search Space Partition with Application to Continuous Space Optimization

摘要

In view of the slowness and the locality of convergence for Particle Swarm Optimization (PSO) in solving complex continuous space optimization problems. Based the idea of wavelet analysis and quantum algorithm, this paper introduces an improved technique called quantum particle swarm optimization based search space partition (SSPQPSO). In this algorithm, the search space partition method is used to increase search accuracy and inhibit premature problem. The positions of particles are encoded by the probability amplitudes of quantum bit. Each quantum bit contains two probability amplitudes, which accelerate the process of searching. Two typical examples indicate that the novel algorithm possesses several advantages such as less chance of being trapped into premature states, higher accuracy and higher stability. As a result, SSPQPSO can be widely applied to many complex continuous space optimization problems which require higher accuracy.