Plateforme logicielle ouverte pour le développement d'algorithmes de planification des opérations

Authors: Attik, Yassine
Advisor: Gaudreault, JonathanQuimper, Claude-Guy
Abstract: Combinatorial optimization concerns the solving of problems for which the variables take discrete values and on which constraints apply. The set of variables and constraints form the model of the problem. A lot of industrial problems can be represented in this form. A solver is a software that takes as input a model and produces a solution. Constraint programming (CP) is one of the algorithmic techniques that can be used within a solver. In this master’s thesis, we develop a new solver. The primary objective is to rely on an easily modifiable solver in order to add new resolution approaches developed by researchers. Moreover, in order to demonstrate the utility of the solver, we develop an approach using that solver in order to generate alternative loading patterns for a kiln in the forest industry. Finally, in this master’s thesis, we present a new technique for solving some CP problems. The filtering algorithms are triggered according to events that occur when solving the problem. We propose a new event that allows to perform a lazy filtering of the variables. We demonstrate, on a classical combinatorial optimization problem (Balanced Incomplete Block Design), that it gives a better performance while maintaining the same level of filtering when compared with classical events.
Document Type: Mémoire de maîtrise
Issue Date: 2018
Open Access Date: 30 August 2018
Grantor: Université Laval
Collection:Thèses et mémoires

Files in this item:
34158.pdf2.67 MBAdobe PDFView/Open
All documents in CorpusUL are protected by Copyright Act of Canada.