Fast Solution to Large-Scale Multiobjective Optimization Problems using Parallel Ant Colony Optimization in Dynamic Environment

在这项研究中，我们提出了一个解决方案，以现实世界中的送货上门服务的问题，大量的道路存在于城市和道路上的交通随着时间的推移迅速变化。寻找最短旅行时间之旅的方法包括一种混合元启发式算法，该算法将蚁群优化（ACO）与Dijkstra算法相结合，这是一种使用实时交通和预测交通的搜索技术，以及一种使用日本真实道路地图和测量交通的方法。提出了一种基于最大最小蚂蚁系统的混合蚁群算法，并提出了一种提高最大最小蚂蚁系统搜索速度的方法。由于道路交通随时间而变化，因此搜索速率在RWDP中很重要。使用东京市中心地图和历史交通数据的实验结果表明，该方法可以找到比传统方法更好的解决方案。

In this research, we presented a solution to real-world delivery problems for home delivery services where a large number of roads exist in cities and the traffic on the roads rapidly changes with time. The methodology for finding the shortest-travel-time tour includes a hybrid meta-heuristic that combines ant colony optimization (ACO) with Dijkstra algorithm, a search technique that uses both real-time traffic and predicted traffic, and a way to use a real-world road map and measured traffic in Japan. We proposed a hybrid ACO for RWDPs that used a MAX-MIN Ant System (MMAS) and proposed a method to improve the search rate of MMAS. Since traffic on roads changes with time, the search rate is important in RWDPs. Experimental results using a map of central Tokyo and historical traffic data indicate that the proposed method can find a better solution than conventional methods.