A two-step framework for dispatching shared agricultural machinery with time windows

Nowadays, farmers and even governments are faced with increasing yields with limited resources, and the extensive use of agricultural machinery is one of the most efficient methods for that. Agricultural machinery usually charges high prices, and it is economically impractical for small-scale farmers to afford them. The shared agricultural machinery is racing ahead full throttle in the market. Farmers can submit a usage request to a shared agricultural machinery company, and the company would then dispatch their machines to farmers to provide operational service. While this new business mode has shown promising benefits, there are new operational challenges. Agricultural production is a strictly seasonal process, and the yield would be affected by working time. Thus, operators would be assigned continuous tasks caused by the overlap of time windows. With a large number of demands, it is challenging for operators to dispatch shared agricultural machinery with time windows efficiently. This study develops a novel two-step dispatching framework for shared agricultural machinery with time windows. In the first step, a model-based spatiotemporal clustering approach is developed to cluster farmlands according to their location, time windows, and crop strain. The shortest route within each cluster of farmlands is also determined. In the second step, shared agricultural machines are routed across the clusters to minimize the dispatching costs. These two steps are formulated as Mixed Integer Linear Programming models, and a two-step heuristic based on CPLEX is proposed to solve these problems. Numerical experiments are conducted with large-scale data from a real-world shared agricultural machinery company. Our computational experiments demonstrate the efficiency, effectiveness, and practicality of the developed approach.