With the busy life of modern people, more and more consumers are preferring to shop online. This change on shopping behavior results in large volumes of packages must be transported, and thus research on logistics planning considering real constraints has increased. To solve this problem, several heuristics or evolutionary methods with expert knowledge were proposed previously, but they are usually inefficient or need a large amount of memory. In this paper, we propose a hybrid approach called Iterative Logistics Solution Planner (ILSP) for not only quickly finding a nice logistics solution but also iteratively improving the solution quality while meeting the real logistics constraints. ILSP contains two main phases including initial logistics solution generation and iterative logistics solution improvement based on the intelligence and knowledge from domain experts. Several algorithms and strategies are designed in ILSP for package partitioning, route planning and quality improvement. From the view of expert systems, the significance and impact of ILSP are simultaneously taking both computational efficiency and iterative quality improvement based on the expert knowledge into account on logistics planning problem with pickup and delivery. Through the rigorous experimental evaluations of real logistics data, the results demonstrated the excellent performance of ILSP.
All Science Journal Classification (ASJC) codes
- Computer Science Applications
- Artificial Intelligence