Otimização via Branch-and-Bound: Teoria, Algoritmos e Exemplos de Aplicação
Otimização via Branch-and-Bound: Teoria, Algoritmos e Exemplos de Aplicação
DOI: https://doi.org/10.22533/at.ed.654112526022
Palavras-chave: Otimização; Programação Linear Inteira; Branch-and-Bound.
Keywords: Optimization; Integer Linear Programming; Branch-and-Bound.
Abstract: This chapter presents a theoretical and applied analysis of the Branch-and-Bound (B&B) method and its relevance in optimizing electrical engineering systems, with a focus on solving integer linear programming (ILP) problems. B&B is a structured approach that organizes the search for the optimal solution by subdividing the original problem into smaller subproblems, leveraging relaxations, heuristics, and pruning strategies to eliminate infeasible regions of the search space. This process significantly reduces computational effort and enhances solution efficiency. Initially, the theoretical foundations of the method are discussed, including its algorithmic structure, evaluation criteria, and search strategies employed to ensure optimal solutions. Subsequently, a practical application is demonstrated using the LINDO (Linear Interactive and Discrete Optimizer) software, a widely used tool in operations research. To validate the method's effectiveness, a detailed case study is presented, addressing the resolution of a maximization problem in ILP. The study involves constructing the enumeration tree, defining upper and lower bounds, and eliminating infeasible subproblems. The results confirm that the B&B method is a robust and effective approach for solving combinatorial optimization and resource allocation problems, making it widely applicable in engineering, data science, and operations research. Its ability to selectively explore the search space while avoiding exhaustive enumeration makes it essential for obtaining optimal solutions in complex systems.
- Joelson Lopes da Paixão
- Gabriel Henrique Danielsson
- Alzenira da Rosa Abaide