Yazid Mati

A genetic local search algorithm for minimizing total weighted tardiness in the job-shop scheduling problem

This paper considers the job-shop problem with release dates and due dates, with the objective of minimizing the total weighted tardiness. A genetic algorithm is combined with an iterated local search that uses a longest path approach on a disjunctive graph model. A design of experiments approach is employed to calibrate the parameters and operators of the algorithm. Previous studies on genetic algorithms for the job-shop problem point out that these algorithms are highly depended on the way the chromosomes are decoded. In this paper, we show that the efficiency of genetic algorithms does no longer depend on the schedule builder when an iterated local search is used. Computational experiments carried out on instances of the literature show the efficiency of the proposed algorithm.

A general approach for optimizing regular criteria in the job-shop scheduling problem

Even though a very large number of solution methods has been developed for the job-shop scheduling problem, a majority has been designed for the makespan criterion. In this paper, we propose a general approach for optimizing any regular criterion in the job-shop scheduling problem. The approach is a local search method that uses a disjunctive graph model and neighborhoods generated by swapping critical arcs. The connectivity property of the neighborhood structure is proved and a novel efficient method for evaluating moves is presented. Besides its generality, another prominent advantage of the proposed approach is its simple implementation that only requires to tune the range of one parameter. Extensive computational experiments carried out on various criteria (makespan, total weighted flow time, total weighted tardiness, weighted sum of tardy jobs, maximum tardiness) show the efficiency of the proposed approach. Best results were obtained for some problem instances taken from the literature.

Complexity of flowshop scheduling problems with a new blocking constraint

This article deals with makespan minimization in the flowshop scheduling problem under the condition of no intermediate storage between machines. A new blocking constraint met in several industrial problems is introduced, and several complexity results are presented from two to five machines. Some problems with four machines in which the new and the classical blocking constraints are mixed, are polynomial. Problems with only the new blocking constraint are polynomial for up to three machines. Although the complexity of the problem with four machines is left open, several cases are shown to be polynomial. Finally the problem with five machines is NP-hard.

Minimizing the makespan in the non-preemptive job-shop scheduling with limited machine availability

This paper addresses the makespan minimization in a job-shop environment where the machines are not available during the whole planning horizon. The disjunctive graph model is used to represent the schedules and the concept of blocks is generalized to include the unavailability periods of machines. To solve the problem, we develop a taboo thresholding heuristic that uses a new block-based neighborhood function. Some sufficient conditions to eliminate the evaluation of non-improving moves are proposed. Experiments performed on existing problem instances of the literature show the efficiency of the proposed heuristic.

Modelling and solving a practical flexible job-shop scheduling problem with blocking constraints

This paper presents a study of a practical job-shop scheduling problem modelled and solved when helping a company to design a new production workshop. The main characteristics of the problem are that some resources are flexible, and blocking constraints have to be taken into account. The problem and the motivation for solving it are detailed. The modelling of the problem and the proposed resolution approach, a genetic algorithm, are described. Numerical experiments using real data are presented and analysed. We also show how these results were used to support choices in the design of the workshop.

Multiresource Shop Scheduling With Resource Flexibility and Blocking

This paper proposes a general scheduling model that extends job-shop scheduling models to incorporate important features of real manufacturing systems. More precisely, each operation can be performed in different modes and requires a different set of resources depending on the mode. Further, we consider blocking constraints that requires to hold resources used for an operation till resources needed for the next operation of the same job are available. A shortest path approach extending the classical geometric approach is proposed for the two-job case. A greedy heuristic is then proposed to schedule N jobs by considering jobs sequentially, grouping scheduled jobs into a combined job and then scheduling it and the next unscheduled job using the shortest path approach. A metaheuristic is then used to identify effective job sequences. Extensive numerical experimentation proves the efficiency of our approach.

A GENETIC-SEARCH-GUIDED GREEDY ALGORITHM FOR MULTI-RESOURCE SHOP SCHEDULING WITH RESOURCE FLEXIBILITY

The Multi-Resource Job-Shop Problem with resource Flexibility (MJSPF) provides a framework for realistic modeling of a wide range of problems encountered in manufacturing systems. The problem is a generalization of the classical job shop problem. Each operation may require a combination of more than one resource and there may be several feasible resource combinations for each operation. The scheduling problem consists in both assigning resources to operations and sequencing operations on the selected resources in order to minimize the makespan. In this paper, a polynomial algorithm for solving a special case with two jobs is proposed, and the concept of a combined job is introduced. Building on these results, a greedy heuristic that considers jobs sequentially according to a given job sequence is proposed for scheduling any number of jobs. The greedy heuristic is guided by a genetic algorithm in order to identify effective job sequences. Computational results on benchmark instances for special cases of the MJSPF show that the general method is competitive with respect to the best known heuristic approaches dedicated to these special cases.

Scheduling challenges and approaches in semiconductor manufacturing

Abstract The semiconductor industry has probably the most complex manufacturing process. The challenges encountered are technological, but concern as well planning and scheduling decision problems. After a short description of the major semiconductor manufacturing steps, this paper concentrates on describing the related scheduling problems. We survey the methods developed in the literature to solve these problems and try to define future interesting research perspectives.

A Reduced-Complexity Algorithm for Two-Job Shop Scheduling Problems with Availability Constraints

This paper addresses the complexity of scheduling problems considering two jobs to schedule and availability constraints imposed on the machines. A polynomial algorithm called temporized geometric approach is first proposed for the minimization of the makespan, under the non-preemption constraint. Then, a generalization to the preemptive case is developed. These algorithms are extensions of the geometric approach, which allows solving the classical two-job scheduling problem.

A GENETIC LOCAL SEARCH ALGORITHM FOR MINIMIZING THE TOTAL WEIGHTED TARDINESS IN A JOB-SHOP

Abstract We consider a job-shop scheduling problem with release dates and due dates, with the objective of minimizing the total weighted tardiness. A genetic algorithm is combined with an iterated local search that uses a neighborhood based on a disjunctive graph model. In this paper, we show that the efficiency of genetic algorithms does no longer depend on the schedule builder when an iterated local search is used. Computational experiments carried out on instances of the literature show the efficiency of the proposed algorithm over the existing methods.