Formulações e heurísticas para o problema de localização de coberturas com sobreposições

• Ano de defesa: 2020
• Autor(a) principal: Araujo, Eliseu Junio [UNIFESP]
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de São Paulo (UNIFESP)
• Palavras-chave em Português: Metaheuristic; Combinatorial Optimization; Hybrid Methods; Coverage Location Problem; Metaheurística; Otimização Combinatória; Método Híbrido; Controle De Sobreposições; Problemas De Localização De Coberturas
• Link de acesso: https://sucupira.capes.gov.br/sucupira/public/consultas/coleta/trabalhoConclusao/viewTrabalhoConclusao.jsf?popup=true&id_trabalho=10889463; https://hdl.handle.net/11600/64930

Coverage location problems and its variations aim to locate facilities strategically to meet the maximum number of demand points that require attendance, meeting different criteria, for example, meeting service capacity constraints. We observe that in many contexts, there are regions that need different priorities for their population concentration, since a region may have different population concentrations. Thus, the priorities are represented by the amount and the organization of coverage zones that must offer as adequate support according to the number of demands covered, avoiding overload of service. Due to the literature have gaps of efficient approaches about this theme, this thesis proposes the overlaps control between coverage zones to meet different prioritization criteria. Thus, a constraint that enables this control is adapted to classical coverage problems. Basically, this constraint quantifies the proportion of overlaps between coverage zones and, depending on the user’s choice, maximizes or minimizes them. Therefore, the control is integrated into the Maximal Coverage Location Problem (MCLP), to the Probabilistic Maximal Coverage location-allocation Problem (PMCLP), and to the Coverage Location Problem (CLP). In instances that optimal solution was not found by CPLEX, the method Density Clustering Search (DCS) was chose to give good solutions to the classical problems due to having re- turned good solutions to problems related to coverage. Results show that overlaps control combination with coverage location problems was efficient, giving the expected solutions. DCS results also show that the method is efficient in terms of objective function value and computational time.