Tipo:	conferenceObject
Título:	A Genetic Algorithm Solution for Combinatorial Problems - The Nuclear Core Reload Example
Autor(es):	SCHIRRU, Roberto PEREIRA, Cláudio Márcio do Nascimento Abreu CHAPOT, J.L.C. SILVA, F.C.
Resumo:	This paper presents one solution of the Traveling Salesman Problem (TSP) based upon genetic algorithms (GA), using the classic crossover, but avoiding the feasibility problem in offspring individuals, allowing the natural evolution of the GA without introduction of heuristics in the genetics crossover operator. The genetic model here presented, that we call the List Model (LM) is based on the encoding and decoding the genotype in the way to always generate a phenotype that has a valid structure, over which will be applied the fitness, represented by the total distance. The main purpose of this work was to develop the basis for a new genetic model to be used in the reload of nuclear core of pressurized water reactors (PWR). In a generic view, this problem can be interpreted as a search of the optimal combination of N different fuel elements in N nuclear core “holes”, where each combination, or load pattern, determines the neutron flux shape and its associate peak factor. The goal is to find out the load pattern that minimizes the peak factor and consequently maximize the useful life of the nuclear fuel. The GA with the List Model was applied to the Angra-1 PWR reload problem and the results are remarkably better than the ones used in the last fuel cycle.
Palavras-chave:	Genetic Algorithm Solution Traveling Salesman Problem Classic Crossover
Idioma:	eng
País:	Brasil
Editor:	Instituto de Engenharia Nuclear
Sigla da Instituição:	IEN
Tipo de Acesso:	openAccess
URI:	http://carpedien.ien.gov.br:8080/handle/ien/1706
Data do documento:	1997
Aparece nas coleções:	Desenvolvimento de Tecnologia para Sistemas Complexos - Trabalhos de Congresso

Os itens no repositório estão protegidos por copyright, com todos os direitos reservados, salvo quando é indicado o contrário.