开放式关联运输调度问题的研究

针对带软时间窗的开放式关联运输调度问题（ Open Incident Vehicle Routing Problem with Soft Time Windows， OIVRPSTW），联系实际应用中连锁店超市中货物供应的车辆路径问题，介绍了粒子群算法的基本原理，采用一种非线性动态自适应调节惯性权重因子的方法，使得惯性系数会随着粒子目标值的变化而自动改变，结合混沌搜索产生初始种群的方法，设计了一种混沌粒子群优化算法。同时也设计了引入了模拟退火机制的混沌遗传算法，自适应地调整交叉概率和变异概率。并用这两种算法来对OIVRPSTW求解，然后与基本的遗传算法求解此模型的结果相比较。实例证明用混沌搜索产生初始种群的方法在求解此类开放式关联运输调度问题是可行的，能取得令人满意的效果。

Research on the Open Incident Vehicle Routing Problems

Aiming at Open Incident Vehicle Routing Problem with Soft Time Windows （ OIVRPSTW） and relating to the VRP（ Vehicle Routing Problem ） of goods supply in chain stores , this paper introduces the fundamental principle of PSO （ Particle Swarm Optimization ） , adopting the nonlinear dynamic self -adapting method to adjust inertia weight factor , so that the weight factor can change along with the target value , combined with the advantage of using chaos to generate initial population and design a kind of CPSO（Chaos Particle Swarm Optimizaton）.Meanwhile, SA （simulated annealing） mechanism is brought in the CGA （Chaos Genetic Algorithm）, adjusting crossover probability and mutation probability adaptively .At last, both of the modified algorithms are applied to solve OIVRPSTW , so does GA （ Genetic Algorithm ） .The result shows that the advantage of using chaos to generate ini-tial population is flexible to solve this kind of OIVRP and can obtain the satisfied result .