Quantum-behaved Particle Swarm Optimization Algorithm Based on Dynamic Dual-population Joint-search Mechanism

摘要

Quantum particle swarm optimization (QPSO) has disadvantages such as rapid loss of species diversity and inability to jump out of local optimum value in the later stage. In this paper, a QPSO algorithm based on dynamic dual-population joint-search mechanism (DJ-QPSO) is proposed. This algorithm establishes two local attraction points in the search area to guide the particle search in the population, and adjusts the global exploration and local exploitation ability by changing the population diversity. Then, the algorithm uses a periodic dynamic-sharing strategy to enable information exchange between the two subgroups. Finally, a global convergence formula is introduced to the search in the later stage to improve algorithm precision. The simulation results of 15 benchmark functions demonstrate that the improved algorithm performs better than comparable algorithms and can effectively deal with complex optimization problems.