Skip to Main content Skip to Navigation
Theses

Programmation linéaire en nombres entiers pour l'ordonnancement cyclique sous contraintes de ressources

Maria Ayala 1
1 LAAS-MOGISA
LAAS - Laboratoire d'analyse et d'architecture des systèmes
Abstract : The resource-constrained modulo scheduling problem (RCMSP) is a general periodic cyclic scheduling problem, abstracted from the problem solved by compilers when optimizing inner loops at instruction level for very long instruction word parallel processors. Since solving the instruction scheduling problem at compilation phase in less time critical than for real time scheduling, integer linear programming (ILP) is a relevant technique for the RCMSP. In this work, we are interested in the methods based on the integer linear programming for the RCMSP. At first, we present a study of the two classic integer linear formulation for the RCMSP. A theoretical evidence of the equivalence between the classic formulations is shown in terms of linear programming (LP) relaxation. Secondly, based on the ILP formulations for the RCMSP, stronger formulations for the RCMSP derived from Dantzig-Wolfe decomposition are presented. In these formulations, the number of variables can be huge, for this reason, we proposed a column generation scheme to solve their LP relaxations. We propose also the heuristics methods based on the Lagragian relaxation and decomposed software pipelining. The heuristic methods search the transformation of the classic integer linear programming for the RCMSP for the performance improvement in the time for the search of solutions. All formulations are compared experimentally on problem instances generated from real data issued from the STMicroelectronics ST200 VLIW processor family.
Document type :
Theses
Complete list of metadatas

https://tel.archives-ouvertes.fr/tel-00604537
Contributor : Arlette Evrard <>
Submitted on : Wednesday, June 29, 2011 - 11:10:16 AM
Last modification on : Friday, January 10, 2020 - 9:10:08 PM
Long-term archiving on: : Friday, September 30, 2011 - 2:22:14 AM

Identifiers

  • HAL Id : tel-00604537, version 1

Citation

Maria Ayala. Programmation linéaire en nombres entiers pour l'ordonnancement cyclique sous contraintes de ressources. Automatique / Robotique. Université Paul Sabatier - Toulouse III, 2011. Français. ⟨tel-00604537⟩

Share

Metrics

Record views

1405

Files downloads

2215