Flexible job-shop scheduling with limited flexible workers using an improved multiobjective discrete teaching-learning based optimization algorithm

Flexible job-shop scheduling problem with worker flexibility (FJSPW) has been frequently investigated during the last decade. Many real-world industries proceed with limited multiskilled workers to optimize cost and improve efficiency, and also to overcome the occasional shortage of workers. Owing to the limited flexible workforce, an excessive workload amongst workers can cause stress, exertion, and lack of focus, leading to low productivity. However, such problems caused due to the limited number of workers are generally not investigated in depth in FJSPW. To consider workforce relief and productivity simultaneously, this research proposes the flexible job-shop scheduling problem with limited flexible workers (FJSPLFW) with the objectives to minimize the makespan, maximum worker workload and total workload of machines. An improved multiobjective discrete teaching-learning based optimization (IMDTLBO) algorithm is introduced to solve the FJSPLFW. A teaching-learning based random partial updating of the learners is proposed to enhance the diversity of learners (population) and avoid local optimum. The number of chapters (elements) of the learner to be updated is considered a key parameter and its value is evaluated with other key parameters through the Taguchi method. The IMDTLBO algorithm is tested on FJSPLFW through various experiments based on 30 newly constructed benchmark instances. The results show that the IMDTLBO algorithm is competitive in solving FJSPLFW.