Improvement on Parallel AQPSO Using the Best Position
Yan Ma, Yang Liu, Deyun Yang, Yuping Chen · 2009
Quantum-behaved Particle Swarm Optimization (QPSO) is a new Particle Swarm Optimization (PSO) algorithm. Compared with Standard PSO (SPSO), it guarantees that particles converge in global optimum point in probability and this algorithm has better performance and stability. This paper introduces an improved Adaptive QPSO algorithm, puts the parallelisms crude of AQPSO and high speed of computer together, and island model is introduced. Multi-swarm Parallel AQPSO (PAQPSO) Algorithm is reported. The algorithm employs the co-evolution model to avoid pre-maturity and improves global search performance. This approach is tested on several accredited benchmark functions and the experiment results show much advantage of PAQPSO to PSOs, and the running time is also decreased in linear.