Skip to Main content Skip to Navigation

Conditional Scheduling Strategies Using Timed Automata

Abstract : In this thesis we develop a methodology for solving conditional scheduling problems where knowing if a task have to be executed is not known in advance but dynamically. The model used is based on timed automata representing the state space to be explored. The problem is formulated as a game against the environment from which we search for a winning strategy (worst case optimal). In the first part we study the deterministic problem of the task graph scheduling and then we extend the framework to the conditional problem. For each problem we study different types of schedules and strategies in order to reduce the state space search, decompositions into chains are proposed to reduce its size, then we investigate several exact algorithms in order to evaluate their efficiency and from which we derive some good heuristics. Experimental results on sets of benchmarks are presented to evaluate the efficiency of each algorithm and the precision of the proposed heuristics, then we deduce theoretical bounds to show the worst case guarantee of each heuristic.
Document type :
Complete list of metadatas
Contributor : Abdelkarim Aziz Kerbaa <>
Submitted on : Thursday, December 21, 2006 - 3:06:42 PM
Last modification on : Thursday, November 19, 2020 - 3:58:01 PM
Long-term archiving on: : Wednesday, April 7, 2010 - 1:09:55 AM


  • HAL Id : tel-00121655, version 1



Abdelkarim Aziz Kerbaa. Conditional Scheduling Strategies Using Timed Automata. Modeling and Simulation. Université Joseph-Fourier - Grenoble I, 2006. English. ⟨tel-00121655⟩



Record views


Files downloads