Resource-constrained Modulo scheduling

Benoît de Dinechin (1) , Christian Artigues (2) , Sadia Azem (3, 4)

1 STMicroelectronics
2 LAAS - Laboratoire d'analyse et d'architecture des systèmes
3 FSTC - Faculté des Sciences, de la Technologie et de la Communication
4 SFL-ENSMSE - Département Sciences de la Fabrication et Logistique

This chapter presents an unusual application of the RCPSP, namely an instruction scheduling problem for very long instruction word processors. The specific characteristics of these processors and the cyclic nature of the problem yields a modulo RCPSP with minimum and maximum time lags. New integer linear programming formulations are proposed and an efficient large neighborhood search technique is used to solve near-optimally large industrial problems in a few minutes.

Mots clés: Resource-constrained project scheduling; Modulo Scheduling; VLIW processors; instruction scheduling; integer linear programming; large neighborhood search

Domaines: Recherche opérationnelle [math.OC]

Benoît de Dinechin, Christian Artigues, Sadia Azem. Resource-constrained Modulo scheduling. Christian Artigues, Sophie Demassey and Emmanuel Néron. Resource Constrained Project Scheduling. Models, Algorithms, Extensions and Applications., ISTE-WILEY, pp.267-277, 2008, Control Systems, Robotics and Manufacturing Series. ⟨hal-00271651⟩