Please use this identifier to cite or link to this item:
|Title:||An agent-based evolutionary approach for manufacturing system layout design|
|Authors:||Pereira, Nuno Emanuel Nunes|
|Publisher:||Faculdade de Ciências e Tecnologia|
|Abstract:||In this thesis it is presented an approach to the problem of layout design for a manufacturing system, which is an important part of its design stage, given that it has influence in the system efficiency and, therefore, in its output rate and fault handling capabilities. The presented approach is based on a Genetic Algorithm (GA) that, by using information provided by the the user through an ontology file, and by using algorithms from graph-theory, designs the layout of a manufacturing system. The instances of the ontology represent manufacturing resources and their characteristics that, when they are being used by the algorithm, are encoded in chromosomes and in their genes. The algorithm begins with a number of chromosomes with low fitness which, with the directed evolution provided by the algorithm, that is restricted by the control parameters that might be tunned by the user, improve with the passing of the new generations. It is considered that the fittest solution is the one that connects, in order, all the resources required by the manufacturing plan, described in the ontology, without the occurrence of overlaps when the layout is constructed. The configuration presented by the transport system that handles parts and materials, in the selected layout, is only dependent on the available resources and on the fitness function used by the GA, being that the last cannot be changed by the user. This approach differs from others by positioning simultaneously all the components of the manufacturing system and not only workstations or transport system. The solution is directed to evolvable assembly systems, purpose for which it was implemented inside an agent, so it can be integrated in a Multiagent System (MAS) to be used in the control of a manufacturing system with minimal changes. Keywords: layout design, manufacturing system, multiagent system, ontology, genetic algorithm.|
|Description:||Dissertação apresentada na Faculdade de Ciências e Tecnologia da Universidade Nova de Lisboa para obtenção do grau de Mestre em Engenharia Electrotécnica e de Computadores|
|Appears in Collections:||FCT: DEE - Dissertações de Mestrado|
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.