Métodos metaheurísticos aplicados a um modelo de planejamento de culturas
Carregando...
Data
2013-03-21
Autores
Oliveira, Cíntia Pimentel de [UNESP]
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
No contexto atual de sustentabilidade, formas alternativas à utilização de produtos que podem levar o planeta a uma crise ambiental estão sendo fortemente estudadas. Na agricultura, uma preocupação brasileira ´e com a redução do consumo de agrotóxicos, já que o país é um dos maiores consumidores mundiais do produto. Neste sentido, o planejamento das atividades agrícolas é uma ação preventiva que pode colaborar com a redução da proliferação de pragas entre as culturas plantadas em lotes e, consequentemente, a redução no consumo de produtos químicos. Ele visa determinar a melhor forma de ocupação de áreas de plantio de modo que a disposição das culturas e/ou variedades em talhões favoreça o controle eficiente de pragas. Neste trabalho é proposto um modelo de otimização 0-1 para minimizar a probabilidade de proliferação de pragas entre as culturas plantadas, respeitando restrições de demanda, ocupação de lotes, período de cultivo das culturas e tempo de planejamento. Nesta modelagem, considera-se uma área de plantio genérica, com lotes irregulares e culturas já plantadas em fazendas vizinhas. Para resolução do modelo foram utilizadas as estratégias Metaheurísticas Algoritmo Genético (AG), Simulated Annealing (SA) e a abordagem híbrida Algoritmo Genético com Simulated Annealing (AG+SA) para obtenção de boas estimativas para o planejamento otimizado das atividades agrícolas. Aplicações práticas a duas instâncias distintas foram realizadas e as melhores soluções factíveis encontradas pelos algoritmos são apresentadas, bem como comparação e discussão dos resultados e desempenho computacional dos métodos. Os resultados computacionais indicam que os algoritmos metaheurísticos propostos encontraram soluções factíveis de boa qualidade em um tempo computacional reduzido, e em especial, as estratégias híbridas demonstraram ser excelentes ferramentas para auxilio no planejamento de plantio
In the sustainability context we are living nowadays, alternative forms to the utilization of products that can lead the Earth to an environmental crisis have been strongly studied. In agriculture, the Brazilian concern is about the reduction of pesticide consumption, once the country is one of the world’s major consumers of it. In this sense, the planning of the agricultural activity is a preventive action that can contribute to reduce the proliferation of pests among crops planted in lots and, consequently, decrease the excessive consumption of such chemicals. This approach aims to determine the best way to use the planting areas arranging the crops and/or varieties in support to an efficient pests control. The research’s objective is to propose an optimization model 0-1 in order to minimize the likelihood of pests proliferation among planted crops, considering constraints demand, lots occupation, crops growing period and planning time. In this model, it is considered a generic planting area with irregular lots and already planted crops in neighboring farms. In the model’s solutions, Genetic Algorithm and Simulated Annealing Metaheuristics strategies were used, as well as a Genetic Algorithm jointly with Simulated Annealing hybrid approach in order to obtain good estimates for the optimized planning of agricultural activities. A practical application for the two instances was performed and the best feasible solutions found by the algorithms are presented as well as a comparison and discussion of the results and performance of computational methods. The computational results indicate that the metaheuristic algorithms found feasible solutions with good quality in a reduced computational time, and in particular, the proposed hybrid strategies have proven to be excellent tools to help planning the planting
In the sustainability context we are living nowadays, alternative forms to the utilization of products that can lead the Earth to an environmental crisis have been strongly studied. In agriculture, the Brazilian concern is about the reduction of pesticide consumption, once the country is one of the world’s major consumers of it. In this sense, the planning of the agricultural activity is a preventive action that can contribute to reduce the proliferation of pests among crops planted in lots and, consequently, decrease the excessive consumption of such chemicals. This approach aims to determine the best way to use the planting areas arranging the crops and/or varieties in support to an efficient pests control. The research’s objective is to propose an optimization model 0-1 in order to minimize the likelihood of pests proliferation among planted crops, considering constraints demand, lots occupation, crops growing period and planning time. In this model, it is considered a generic planting area with irregular lots and already planted crops in neighboring farms. In the model’s solutions, Genetic Algorithm and Simulated Annealing Metaheuristics strategies were used, as well as a Genetic Algorithm jointly with Simulated Annealing hybrid approach in order to obtain good estimates for the optimized planning of agricultural activities. A practical application for the two instances was performed and the best feasible solutions found by the algorithms are presented as well as a comparison and discussion of the results and performance of computational methods. The computational results indicate that the metaheuristic algorithms found feasible solutions with good quality in a reduced computational time, and in particular, the proposed hybrid strategies have proven to be excellent tools to help planning the planting
Descrição
Palavras-chave
Biometria, Pragas agricolas - Controle biologico, Produtos químicos agrícolas, Estatistica matematica, Agricultural pests - Biological control
Como citar
OLIVEIRA, Cíntia Pimentel de. Métodos metaheurísticos aplicados a um modelo de planejamento de culturas. 2013. 90 f. Dissertação (mestrado) - Universidade Estadual Paulista Julio de Mesquita Filho, Instituto de Biociências de Botucatu, 2013.