A Multi-Load AGV Task Allocation Method Considering Battery Constraints

The multi-robot task allocation problem is a critical component of intelligent material delivery in manufacturing workshops. This study addresses the task allocation problem for multiload AGVs (Automated Guided Vehicles) with the optimization objectives of minimizing the number of AGVs used, the total path time cost, and the time cost difference among AGVs. A multi-objective optimization mathematical model is developed, which comprehensively considers AGV battery levels and load capacities. To solve this model, an improved NSGA-II algorithm is proposed. The algorithm employs a combined partially matched crossover (CPMX) strategy to better explore the search space and improve the quality of solutions, while ensuring that the offspring generated during the crossover process satisfy the sequential constraints of pick-up and delivery nodes. Additionally, to generate an effective and constraint-compliant initial population, a novel chromosome encoding structure incorporating charging points is designed, and a strategy-based two-stage heuristic path generation method is introduced. During the mutation phase, three mutation operators are used to introduce new features into the population and expand the search space for feasible solutions. Finally, simulation experiments with real-world data demonstrate that the improved NSGA-II algorithm exhibits superior optimization capability and iterative efficiency compared to the basic NSGA-II algorithm and similar multi-objective algorithms.