Multi-robot Task Allocation Using Island Model Genetic Algorithm

Abstract

Mobile robots have been used in several areas as agriculture, vigilance, and indoor transportation in factories and warehouses. The warehouses’ efficiency is directly related to transportation in the order picking process. Using mobile robots can improve transportation efficiency when associated with optimized task allocation. This paper shows a method, using the island model genetic algorithm, to allocate tasks with deadlines in multi-robot systems where mobile robots have different payloads and speed capacities. The approach’s performance was evaluated by varying parameters of the algorithm and the size of the task set, and the robot set involved in the allocation process. The performance criteria adopted were success rate in allocating all tasks, percentage of assigned tasks in non-complete solutions, and obtaining a complete solution. The results have shown that the method brings a more significant number of complete solutions in sets with a small and medium number of tasks. Besides that, the solutions obtained meets deadlines and show energy consumption reduction.