Renewable energy is an alternative to non-renewable energy to reduce the carbon footprint of manufacturing systems. Finding out how to make an alternative energy-efficient scheduling solution when renewable and non-renewable energy drives production is of great importance. In this paper, a multi-objective flexible flow shop scheduling problem that considers variable processing time due to renewable energy (MFFSP-VPTRE) is studied. First, the optimization model of the MFFSP-VPTRE is formulated considering the periodicity of renewable energy and the limitations of energy storage capacity. Then, a hybrid non-dominated sorting genetic algorithm with variable local search (HNSGA-II) is proposed to solve the MFFSP-VPTRE. An operation and machine-based encoding method is employed. A low-carbon scheduling algorithm is presented. Besides the crossover and mutation, a variable local search is used to improve the offspring’s Pareto set. The offspring and the parents are combined and those that dominate more are selected to continue evolving. Finally, two groups of experiments are carried out. The results show that the low-carbon scheduling algorithm can effectively reduce the carbon footprint under the premise of makespan optimization and the HNSGA-II outperforms the traditional NSGA-II and can solve the MFFSP-VPTRE effectively and efficiently.
Considering the increasing concern on sustainable development from manufacturers, focus is given to three kinds of indicators of sustainable development, that is, economy, environment, and society, and schedule two types of resource, that is machines and workers, simultaneously in the classical flexible job shop scheduling problem. The authors define it as a sustainable flexible job shop scheduling problem considering dual resources (SFJSPCDR). First, a model of the SFJSPCDR is formulated to optimise the makespan, the energy consumption, and the ergonomic risk simultaneously. Second, an improved survival duration‐guided NSGA‐III algorithm (SDG‐NSGA‐III) is proposed to solve SFJSPCDR. The survival duration of each individual determines whether it takes part in generating offspring. In order to balance the energy consumption and ergonomic risk while minimising the makespan, a double‐low decoding algorithm is proposed, which is composed of two decoding algorithms. Cross‐generation selection is employed with the non‐dominated sorting, and the next‐generation population is selected according to the reference point‐based selection strategy. In addition, a restart strategy is also integrated to improve the exploration and exploitation performance of the SDG‐NSGA‐III algorithm. Finally, a group of experiments are carried out and the results prove the effectiveness of the proposed algorithm.
The operating room is one of the most capital-intensive resources for a hospital. To achieve further improvements and to restrict cost increases, hospitals may need to operate more efficiently with the resources they already possess. The paper considers the joint problem of planning and scheduling patients in operating rooms on an operational level (weekly basis) with two objectives: maximizing the overall patients' satisfaction and minimizing the cost of overtime in operating rooms as well as the daily cost of operating rooms and recovery beds, which is NP-hard. The decision problem is solved using a bi-layer discrete particle swarm optimization, introducing a repair mechanism for infeasible solutions, specific operators like crossover, insertion and exchange. Moreover, a gap finding scheduling heuristic is designed to solve the surgical case sequencing problem. We first compare the performance of the proposed solution method to that of Fei et al. for three instances separately, using data of a Chinese hospital. Next, the efficient Pareto solutions for the joint problem are presented. The results show that the bi-layer discrete particle swarm optimization can solve the operating room scheduling efficiently and effectively.