Modèles déterministe, stochastique et multicritère pour l'équilibrage de lignes d'assemblage

Abstract : Our work considers the problem of balancing the assembly line (ALBP). This is a combinatorial optimization problem which consists in assigning operations to stations of the assembly line while respecting various constraints in order to optimize a criterion of efficiency. Two types of problems are defined, based on the objective to minimize. The type I problem (SALBP-1) minimizes the number of stations in a given cycle time. And the problem of type II (SALBP-2) minimizes the cycle time given by the largest time station with a given number of stations. Problems of type II are only considered in our work. Firstly, a method for determining the lower bound of the cycle time which respects the number of stations is determined. This method combines the Lagrangian relaxation and the columns generation method. The Lagrangian relaxation is used to relax the constraints of precedence. The problem of the Lagrangian relaxation is given by the columns generation. Then, a new heuristic is proposed for solving the SALBP-2 problem. This heuristic consists of two steps. In the first step, an initial solution is produced by an heuristic based on a weight corresponding to the position of operations and a threshold defined for each station. The, in the second step, the solution is improved by a process of transfer and exchange. Finally, meta-heuristics are used to solve deterministic and stochastic problems.Two main meta-heuristics are considered: the electromagnetism-like mechanism algorithm (EM) and the estimated distribution (ED). The performance of our method is confirmed via simulation and it is compared with simulated annealing (SA). Due to its better performance, EM was chosen to balance the stochastic lines with random operation times. In this case, the cycle time is minimized under constraint of the reliability of the line which must be greater than a given value. EM is also used to solve multi-objective problems such as minimizing the cycle time and maximizing the reliability of the line by determining a set of Pareto-optimal solutions
Document type :
Theses
Complete list of metadatas

Cited literature [46 references]  Display  Hide  Download

https://hal.univ-lorraine.fr/tel-01752564
Contributor : Thèses Ul <>
Submitted on : Thursday, March 29, 2018 - 1:51:18 PM
Last modification on : Tuesday, April 10, 2018 - 1:27:59 AM

File

Gu.Liya.SMZ0803.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : tel-01752564, version 1

Collections

Citation

Liya Gu. Modèles déterministe, stochastique et multicritère pour l'équilibrage de lignes d'assemblage. Autre. Université Paul Verlaine - Metz, 2008. Français. ⟨NNT : 2008METZ003S⟩. ⟨tel-01752564⟩

Share

Metrics

Record views

55

Files downloads

572