整车物流网络规划问题的混合粒子群算法研究

Research on Hybrid Particle Swarm Optimization for Automobile Logistics Network Design Problem

综合考虑整车物流系统中的运输规模经济效应、库存控制策略、设施、服务质量等决策因素,建立了整车物流网络规划集成优化模型.给出了一种流预测算法和粒子群算法相结合的求解方法,用粒子群算法搜索物流网络可行结构,用流预测算法确定其最优运输路径,二者相互协调实现最优解的搜索.在粒子群搜索过程还加入了交叉变异操作来增加种群的多样性,以避免早熟收敛.实例仿真表明混合粒子群算法的运行效率有显著提高,且有更高概率搜索到全局最优.

According to the practical operation characteristic of automobile logistics network, the integrated optimization model is presented, which provides an integrated view of transportation economies- of- scale, inventory and facility costs as well as service quality. The solution combined the flow prediction algorithm and particle swarm optimization （PSO） is presented. In this solution, PSO is used to search feasible structure of logistics network, while flow prediction algorithm is used to decide its optimal transportation route. Evolution operation such as crossover and mutation is also embedded to avoid the common defect of premature convergence. Simulations are given to confirm this hybrid particle swarm optimization work efficiency and the probability of finding the global optimal value are enhanced.