Skip to Main content Skip to Navigation
Theses

Ordonnancement multi-objectif d'ateliers complexes de type job-shop : application à la fabrication de semiconducteurs

Abstract : This work deals with a real-life complex scheduling problem arising in semiconductor manufacturing where dispatching rules are still widely used. Optimization algorithms are a promising alternative to dispatching rules, provided that the solved problem encompasses the rich set of complex constraints and criteria. We consider a flexible job-shop scheduling problem with p-batching, reentrant flows, sequence-dependent setup times, unavailability periods, time lags and release dates. Different criteria must be considered to optimize the different operational performances: Overall throughput, target satisfaction, machine utilization and cycle time.The proposed heuristic approach relies on the adaptation of the disjunctive graph that was introduced in a previous thesis, called batch-oblivious where batching decisions are encoded in the arc weights. This graph is extended to allow the modeling of the internal resources of complex batching machines. An efficient algorithm is proposed to simultaneously compute start times and improve the solution during the graph traversal by filling underutilized batches. In addition to this integrated algorithm, the solution is improved within a simulated annealing metaheuristic. Depending on whether the preferences of the decision-maker are given before the search process, different approaches to handle the multiobjective aspect of the problem are studied and compared. The different components are embedded within a parallelized implementation of the GRASP metaheuristic. Different experiments on large size industrial instances show the significant improvement that can be brought by the proposed approach in computational times of several minutes.
Document type :
Theses
Complete list of metadatas

Cited literature [158 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-02884923
Contributor : Abes Star :  Contact
Submitted on : Tuesday, June 30, 2020 - 11:34:17 AM
Last modification on : Friday, October 23, 2020 - 5:02:27 PM

File

Tamssaouet-karim-diff.pdf
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-02884923, version 1

Collections

Citation

Karim Tamssaouet. Ordonnancement multi-objectif d'ateliers complexes de type job-shop : application à la fabrication de semiconducteurs. Other. Université de Lyon, 2019. English. ⟨NNT : 2019LYSEM016⟩. ⟨tel-02884923⟩

Share

Metrics

Record views

61

Files downloads

48