Two Deadline Reduction Algorithms for Scheduling Dependent Tasks on Parallel Processors

• Published in: Integration of Constraint Programming, Artificial Intelligence, and Operations Research Vol. 12735; pp. 214 - 230
• Main Authors: Hanen, Claire, Kordon, Alix Munier, Pedersen, Theo
• Format: Book Chapter
• Language: English
• Published: Switzerland Springer International Publishing AG 2021; Springer International Publishing
• Series: Lecture Notes in Computer Science
• Subjects: Energetic reasoning; Precedence constraints; Preemptive relaxation; Scheduling problem
• ISBN: 9783030782290; 3030782298
• ISSN: 0302-9743; 1611-3349
• DOI: 10.1007/978-3-030-78230-6_14

This paper proposes two deadline adjustment techniques for scheduling non preemptive tasks subject to precedence relations, release dates and deadlines on a limited number of processors. This decision problem is denoted by P|prec,ri,di|⋆ $$P\vert prec, r_i, d_i\vert \star $$ in standard notations. The first technique is an extension of the Garey and Johnson algorithm that integrates precedence relations in energetic reasoning. The second one is an extension of the Leung, Palem and Pnueli algorithm that builds iteratively relaxed preemptive schedules to adjust deadlines. The implementation of the two classes of algorithms is discussed and compared on randomly generated instances. We show that the adjustments obtained are slightly different but equivalent using several metrics. However, the time performance of the extended Leung, Palem and Pnueli algorithm is much better than that of the extended Garey and Johnson ones.