基于非对称势阱的量子粒子群算法及其应用

Quantum Particle Swarm Optimization Based on Asymmetric Quantum Potential Well and Application

针对量子粒子群算法在处理自变量具有有限定义域的问题时易陷入局部最优解的问题,对算法的量子模型加以改进,提出了基于非对称势的量子粒子群算法(asymmetric potential well based quantum particle swarm optimization,AQPSO)。该算法认为粒子处于非对称势阱中,势阱的参数由当前的最优位置和自变量的定义域共同决定。而在求解粒子在空间分布的波函数时,又采用了参数消减方法,只需人工指定越限概率,简化算法流程。最后,通过算例验证,该方法的全局搜索能力显著提升,在处理高维、复杂、强干扰性问题时,具有显著优势。

Quantum particle swarm optimization（QPSO） may fall into local optimal solution when solving problems with variables with limited definition domain. In order to solve this problem, quantum particle swarm optimization based on asymmetric quantum potential（AQPSO） is proposed to improve QPSO quantum model. In this method, particles are located in asymmetric potential well. Potential well parameters are determined by current optimal location and definition domain. When solving wave-function indicating particle distribution, a method of reducing parameter number is proposed. This parameter-reducing method simplifies algorithm flow by specifying only one parameter： out-of-limit probability. Simulation results prove that global search performance of the new algorithm is improved significantly with advantage in solving complex problems with high-dimensional variables and strong interference.