Formulações matemáticas e algoritmo híbrido para o problema de programação de sessões técnicas de conferências

• Ano de defesa: 2019
• Autor(a) principal: Correia, Rubens Godofredo Soares
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal da Paraíba; Brasil; Informática; Programa de Pós-Graduação em Informática; UFPB
• Palavras-chave em Português: Otimização Combinatória; Conference Scheduling; Algoritmo híbrido; Iterated local Search; Combinatorial Optmization; Hybrid Algorithm; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: https://repositorio.ufpb.br/jspui/handle/123456789/19656

This work addresses the conference scheduling problem of technical sessions. The objective is to maximize the benefit of scheduling papers with common topics in the same session, satisfying resources constraints, as well as those imposed by the organizers. Two mathematical formulations based on integer linear programming are presented for the problem, which in turn is NP-Hard. Given the limit of the models’ scalability, a hybrid algorithm that combines a metaheuristic based on local search and a procedure based on mathematical programming is proposed to solve realistic large size instances. Computational experiments where conducted in small and medium size instances, where the developed heuristic was capable of finding most known optima. The algorithm was compared with manually generated solutions, and it was also used as a tool to generate the schedule of some conferences in Brazil.