Skip to Main content Skip to Navigation
New interface

Structuration de lignes d'usinage : méthodes exactes et heuristiques

Abstract : The doctoral work deals with the preliminary stage of the design of transfer machining lines. To minimise the cost of such lines while respecting a desired productivity, the operations are grouped into blocks. All the operations of a same block are executed simultaneously by the same spindle-head. Spindle-heads of the same workstation are activated sequentially. The grouping of operations into blocks is limited by compatibility constraints of operations, by those due to the possibility to design corresponding spindle-heads and by those due to the compatibility between the spindle-heads to stay on the same workstation. We assume that all the constraints are established beforehand. The problem then consists in finding the grouping of operations into blocks and of blocks on workstations in such way that the weighted sum of workstations and blocks would be the smallest one as possible. We propose a set of new and efficient models and algorithms for this difficult problem. Our methods are based on two main approaches: mixed linear programming and stochastic search in a decision tree. The first approach is used to develop MIP models giving exact solutions for the considered problem but limited by its size; the second approach gives algorithms which do not guaranty the optimality of the solutions but which are faster. We have also combined the two approaches to obtain heuristic algorithms integrating exact models (using them for sub-problems). This coupling is based on the decomposition approach of the initial problem.
Document type :
Complete list of metadata
Contributor : Thèses UL Connect in order to contact the contributor
Submitted on : Thursday, March 29, 2018 - 12:40:33 PM
Last modification on : Tuesday, April 10, 2018 - 1:27:59 AM


Files produced by the author(s)


  • HAL Id : tel-01750207, version 1



Brigitte Finel. Structuration de lignes d'usinage : méthodes exactes et heuristiques. Autre. Université Paul Verlaine - Metz, 2004. Français. ⟨NNT : 2004METZ021S⟩. ⟨tel-01750207⟩



Record views


Files downloads