The no-wait flow shop scheduling problem (NWFSP) plays a crucial role in the allocation of resources in multitudinous industries, including the steel, pharmaceutical, chemical, plastic, electronic, and food processing industries. The NWFSP consists of n jobs that must be processed in m machines in series, and no job is allowed to wait between consecutive operations. This project deals with NWFSP with sequence-dependent setup times for minimizing earliness and tardiness. From the literature review of the last five years in NWFSP, it is noticeable that only around 1.92% of the researchers have studied that multi-objective function, which could help to improve the productivity of industries where methods such as just in time are considered. Besides, there is no information about previous researchers that have solved this problem with sequence-dependent setup times. Firstly, a MILP model is proposed to solve small instances, and secondly, a genetic algorithm (GA) is developed as a solution method for medium and large instances. Compared with the mathematical model for small instances, the GA obtained the optimal solution in 100% of the cases. For medium and large instances, the GA improves in an average of 31.54%, 38.09%, 44.58%, 47.72%, and 37.33% the MDD, EDDP, ATC, SPT, and LPT dispatching rules, respectively.

Job selection and scheduling are among the most important decisions for production planning in today’s manufacturing systems. However, the studies that take into account both problems together are scarce. Given that such problems are strongly NP-hard, this paper presents an approach based on two heuristic algorithms for simultaneous job selection and scheduling. The objective is to select a subset of jobs and schedule them in such a way that the total net profit is maximized. The cost components considered include jobs & apos; processing costs and weighted earliness/tardiness penalties. Two heuristic algorithms; namely scatter search (SS) and simulated annealing (SA), were employed to solve the problem for single machine environments. The algorithms were applied to several examples of different sizes with sequence-dependent setup times. Computational results were compared in terms of quality of solutions and convergence speed. Both algorithms were found to be efficient in solving the problem. While SS could provide solutions with slightly higher quality for large size problems, SA could achieve solutions in a more reasonable computational time.

Skill management is a key factor in improving effectiveness of industrial companies, notably their maintenance services. The problem considered in this paper concerns scheduling of maintenance tasks under resource (maintenance teams) constraints. This problem is generally known as unrelated parallel machine scheduling. We consider the problem with a both objectives of minimizing total weighted tardiness (TWT) and number of tardiness tasks. Our interest is focused particularly on solving this problem under skill constraints, which each resource has a skill level. So, we propose a new efficient heuristic to obtain an approximate solution for this NP-hard problem and demonstrate his effectiveness through computational experiments. This heuristic is designed for implementation in a static maintenance scheduling problem (with unequal release dates, processing times and resource skills), while minimizing objective functions aforementioned.

This paper addresses the stochastic permutation flow shop problem (SPFSP) in which the stochastic parameters are the processing times. This allows the modeling of setups and machine breakdowns. Likewise, it is proposed a multi-objective greedy randomized adaptive search procedure (GRASP) coupled with Monte-Carlo Simulation to obtain expected makespan and expected tardiness. To manage the bi-objective function, a sequential combined method is considered in the construction phase of the meta-heuristic. Moreover, the local Search combines 2-optimal interchanges with a Pareto Archived Evolution Strategy (PAES) to obtain the Pareto front. Also, some Taillard benchmark instances of deterministic permutation flow shop problem were adapted in order to include the variation in processing times. Accordingly, two coefficients of variation (CVs) were tested: one depending on expected processing times values defined as twice the expected processing time of a job, and a fixed value of 0.25. Thus, the computational results on benchmark instances show that the variable CV provided lower values of the expected makespan and tardiness, while the con-stant CV presented higher expected measures. The computational results present insights for further analysis on the behavior of stochastic scheduling problems for a better approach in real-life scenarios at industrial and service systems.