Incorporating human workforce into flexible manufacturing systems is essential for maintaining companies’ competitiveness. Fully automated systems restrict flexibility, whereas the workforce could be considered to improve it : humans can learn, understand, and adapt. This paper studies a flexible manufacturing system that incorporates human operators and considers two variants of a problem : independent resources (two operators assigned to separate stations requiring workforce) and shared resources (one operator assigned to both stations). The objective is to evaluate their impact on system performance, including workforce-related indicators. A deterministic approach based on mathematical models is proposed for small and mid-sized instances, while a simulated annealing algorithm is used for larger instances. Initial solutions for the latter are generated using the NEH heuristic. The computational results show that independent resources lead to unbalanced workloads, whereas shared resources increase the makespan by 12.91% in average.

This paper investigates the permutation flowshop scheduling problem subject to maximum and minimum time lag constraints with the objective of minimizing the makespan, a decision problem challenge frequently encountered in many manufacturing environments. This problem is known to be NP-hard, and a number of constructive heuristics have been proposed in the literature. By leveraging two problem-specific properties (i.e. reversibility and critical path), we propose four new constructive heuristics. The comprehensive computational evaluation carried out demonstrates that the proposed heuristics significantly outperform the existing ones in both solution quality and efficiency.

We study the scheduling of jobs on identical parallel machines billed under a time-of-use electricity tariff, with the goal of minimizing the completion time of the last job while keeping the total energy cost within a fixed budget. The duration of a job shortens with its position on a machine, following a DeJong learning curve. On 60 benchmark instances we compare nine methods: a weak reference rule, five classical heuristics (list scheduling, longest-processing-time assignment, price-aware insertion, simulated annealing, and a genetic algorithm), and three reinforcement-learning policies trained with the learning effect active. When only the completion time matters, the strong heuristics closely rival the learned policies at a fraction of their running time. Under an active energy budget, however, the strong heuristics lose feasibility almost everywhere, and the energy-aware methods, led by the learning-based policies, deliver the lowest feasible completion time.

In this paper, scheduling resiliency is analyzed, defined as the deterioration of the performance measure used in the scheduling process after a disruption occurs and a reaction method has been applied to respond to this disruption. The unrelated parallel machines scheduling problem with makespan objective is considered, and a machine breakdown occurs before the schedule is executed and cannot be repaired. Two reaction methods are compared: Rescheduling all jobs on the remaining machines (named as rescheduling problem); or freezing the scheduled jobs on the operational machines, and only scheduling those assigned to the down machine (named as availability problem). Results show that, in low-variability scenarios the deterioration of the objective function by the availability problem has no statistically significant differences compared to the rescheduling problem, requiring negligible computational times. These findings imply that, in specific scenarios, the frozen policy can be an effective approach for achieving resilience in manufacturing scheduling.