Biased random-key genetic algorithm for cobot assignment in an assembly/disassembly job shop scheduling problem

Abstract

Abstract Nowadays many manufacturing companies try to improve the performance of their processes by including innovative available technologies such as collaborative robots. Collaborative robots are robots where no safety distance is necessary, through cooperation with human workers they can increase production speed. In this paper we consider the collaborative robot assignment combined with the job shop scheduling problem. To solve this problem, we propose a genetic algorithm with a biased random-key encoding. The objective function for the optimization is a weighted function that factors in production cost and makespan that should be minimized. We propose a special encoding of the solution: the assignment of cobots to workstations, the assignment of tasks to different workstations and the priority of tasks. The results show how much the weighted objective function can be decreased by the deployment of additional collaborative robots in a real-world production line. Additionally, the biased random-key encoded results are compared to typical integer encoded solution. With the biased random-key encoding, we were able to find better results than with the standard integer encoding.