Volume 53, Number 2, April-June 2019
|Page(s)||593 - 608|
|Published online||05 June 2019|
Evaluation of a new decision-aid parameter for job shop scheduling under uncertainties
LS2N, Laboratoire des Sciences du Numérique de Nantes, UMR CNRS 6004, Nantes, France
2 MELT, Manufacturing Engineering Laboratory of Tlemcen, Algeria
* Corresponding author: firstname.lastname@example.org
Accepted: 5 October 2017
This paper addresses the groups of permutable operations method. This method is a flexible scheduling approach to hedge against uncertainties and is composed of a proactive/reactive phase. The proactive phase consists of computing a set of solutions (schedule) to a scheduling problem, leaving the choice of executing one of these solutions during the reactive phase according to the current state of the shop floor. During the reactive phase, the remaining decisions have to be made in real-time. The worst-case evaluation of the remaining solutions is a decision-aid parameter used during the reactive phase in order to control the final schedule from exceeding a worst-case performance. While the existing literature only tackles the worst-case evaluation of the groups of permutable operations, this paper deals with its best-case evaluation. For solving this problem, a new lower bound calculating this parameter in polynomial time is proposed. The computational efficiency of this parameter in a reactive algorithm exhibits very good performance. Moreover, the experiments show the robustness of this evaluation allowing this parameter to be used in an unstable job shop environment.
Mathematics Subject Classification: 90B35 / 90B36 / 68M20
Key words: Scheduling / decision-aid system / lower bounds / Job shop / makespan
© EDP Sciences, ROADEF, SMAI 2019
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.