A multi-threaded local search algorithm and computer implementation for the multi-mode, resource-constrained multi-project scheduling problem

•Iterated Variable Neighborhood search for the multi-project, multi-mode resource-constrained project scheduling problem.•Implementation of different parallel computing variants.•Speedup test on parallel, multi-core hardware.•Successful application to the data sets of the MISTA 2013 Challenge.•New b...

Full description

Saved in:
Bibliographic Details
Published in:European journal of operational research Vol. 256; no. 3; pp. 729 - 741
Main Author: Geiger, Martin Josef
Format: Journal Article
Language:English
Published: Amsterdam Elsevier B.V 01.02.2017
Elsevier Sequoia S.A
Subjects:
Iterated local search
Local search
Multi-mode resource-constrained multi-project scheduling
Project management
Scheduling algorithms
Searches
Studies
Variable neighborhood search
Local search
Multi-mode resource-constrained multi-project scheduling
Variable neighborhood search
Iterated local search
ISSN:0377-2217, 1872-6860
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:•Iterated Variable Neighborhood search for the multi-project, multi-mode resource-constrained project scheduling problem.•Implementation of different parallel computing variants.•Speedup test on parallel, multi-core hardware.•Successful application to the data sets of the MISTA 2013 Challenge.•New best known results for 1371 of the 4320 MMLIB50, MMLIB100, MMLIB+ datasets. The article describes a solution approach for the multi-mode, resource-constrained multi-project scheduling problem. Our key ideas are based on the concepts of Variable Neighborhood Search, together with Iterated Local Search. A particular focus of this work is devoted towards the parallel implementation of such local search solution techniques. With the increasing availability of parallel (multi-core) computer hardware, we believe that such techniques are about to gain attractiveness in the future. Experiments have been conducted on benchmark instances of the MISTA 2013 Challenge, as well as on the well-known MMLIB-datasets. The approach described in this article ranked second in the Challenge, and found new best results to 1371 of the 4320 MMLIB-instances. Note that the source-code of our implementation has been made available under http://dx.doi.org/10.17632/cw95t56hjv.1. The software may be freely used for non-profit (research/education) purposes.
Bibliography:SourceType-Scholarly Journals-1
ObjectType-Feature-1
content type line 14
ISSN:0377-2217
1872-6860
DOI:10.1016/j.ejor.2016.07.024