A Self-learning Harris Hawks Optimization Algorithm for Flexible Job Shop Scheduling with Setup Times and Resource Constraints

Abstract

The flexible job shop scheduling problem (FJSP) is a popular research topic in the field of production scheduling. Traditional FJSP ignores sequence-dependent setup times and resource constraints. However, these constraints should be taken into account during the manufacturing process. To deal with the extended FJSP with sequence-dependent setup times and resource constraints, this paper proposes a self-learning harris hawks optimization (SLHHO) algorithm. The goal of this algorithm is to get the smallest makespan. In the proposed algorithm, we use two-vector code to encode machine sequence and operation sequence, design a new decoding method to satisfy the resource constraints, and use reinforcement learning to optimize the key parameters of the algorithm intelligently. We compare it with other three effective algorithms on benchmark instances with varying scales. The experimental result shows that SLHHO performs better and can get a more effective solution.