种群规模自适应粒子群在配送中心连续型选址中的应用

本文依照更具有现实意义的“加工厂—配送中心—用户”的模式建立物流配送中心连续型选址模型,并针对较大规模的选址问题提出求解算法。该算法是将具有较强鲁棒性的自适应粒子算法和改进的ALA(Alert Location-Allocation)方法结合而得,该算法中种群规模自适应变化,对经典粒子移动方程进行改进,消除了学习因子,惯性因子随粒子适应值自适应变化,改进的ALA方法提高了算法计算效率。数值试验表明,本文所建模型具有一定的实践优越性,所提出的算法能有效避免陷入局部最优,寻优能力和鲁棒性均较强。

Application of Swarm Size Adaptive Particle Swarm Optimization to Continuous Location of Distribution Center

In this paper, according to the pattern of “plant-distribution center-users” which is more realistic, a continuous location model of logistics distribution center is established, and a solving algorithm is proposed to solve the large-scale location problem. The algorithm is derived from the combination of improved ALA method with adaptive PSO whose robustness is stronger. In the algorithm, swarm size changes adaptively, the classical particles move equation is improved, the study factors are eliminated, the inertia factor changes adaptively according to fitness value, and the improvement of ALA method increases the algorithm efficiency. Numerical experiments show that the model has practical advantages to a certain extent, and that the algorithm whose optimization ability and robustness are stronger can effectively avoid getting the local optimal.