基于随机游走的分类垃圾回收最优路径规划

Optimization Method of Logistics Paths Planning for Categorical Waste Recycling Based on Random Walk

分类垃圾回收是逆向物流的重要研究内容,物流路径越短意味着回收成本越少。在垃圾分类回收过程中,通过对垃圾的回收路径进行合并可以共享运输资源从而达到节约成本的作用,故本文将垃圾分类回收的路径规划问题假设为多源多目标的路径规划问题,并给出了路径集合中不含重复边的总长度优化模型。当网络规模增长到一定程度时,通过精确计算方法得到模型的最优值几乎是不可能的,为此提出了一种基于随机游走的最优路径集合选取算法。模拟实验验证了该方法的有效性和高效性,与基于Dijkstra算法的最短路径求和算法相比不仅准确性高,而且具有很高的执行效率。

Categorical waste recycling is an important research issue in reverse logistics field, the shorter the logist ics paths are, the lower cost of recycling is. During the processing of categorical waste recycling, the recycling cost could be reduced according to sharing transportation by merging recycling paths of waste. In this paper, we transform the problem of path planning for categorical waste recycling into the problem of path planning for multiple sources and targets, and present a total length optimization model that doesn＇t contain any edge multiple times in the path set. When the scale of network extends to some degree, it is impossible to calculate the accurate optimal resolution of the model. So we propose a random walk based optimal choosing algorithm of path set. The proposed algorithm can reduce the total path length by merging common edges in different paths, and is more accurate and efficient than the Dijkstra based algorithm for summing up all lengths of the shortest paths. Finally, we validate the effective and efEiciency of the proposed algorithm by simulation experiments.