Heuristic-Based Collaborative Multi-Machine Scheduling for Emergency Harvesting in Intensive Farmland

Abstract With the extension of intensive farmland and growing production demand, traditional single agricultural machine farming fails to meet agricultural emergency needs, and time-sensitive harvesting tasks urgently require multiple agricultural machines collaboration. This study aims to design an optimal plan for agricultural machinery fleets to perform crop harvesting. Unlike traditional scheduling and path planning, the emergency scheduling model for intensive farmland proposed in this study targets a unique scenario. This scenario is characterized by agricultural machinery far outnumbering farmland units and multiple machines serving each unit collaboratively. Accordingly, the goal of the model is to minimize makespan (i.e., the completion time of the last machine). Compared to traditional decentralized farm-land scheduling models, this study addresses the challenges of collaborative path planning and resource allocation. It provides an efficient theoretical framework for agricultural machinery collaboration. To efficiently solve the proposed model, the study proposes a two-stage heuristic scheduling algorithm: first overall agricultural machinery allocation, then conduct refined optimization, aiming for rational resource utilization and task completion within limited time. Experimental results obtained using actual farmland data show that the proposed algorithm outperforms representative methods and the Gurobi solver. Notably, this superiority holds across problems of various scales, extending to both solution quality and computational efficiency.