Core Concepts
Biased Random-Key Genetic Algorithms (BRKGA) are a versatile metaheuristic framework that has been widely applied to solve a diverse range of optimization problems, demonstrating its efficiency and adaptability.
Abstract
This systematic literature review and bibliometric analysis examines the existing research on Biased Random-Key Genetic Algorithms (BRKGA). The review covers around 250 papers, providing a comprehensive overview of the various applications of BRKGA, ranging from classical combinatorial optimization problems to real-world industrial scenarios and emerging areas like hyperparameter tuning in machine learning.
The citation analysis identifies the most influential researchers and papers in the field, with Resende, M.G.C. and Gonçalves, J.F. being the most prolific authors. The co-citation analysis reveals the theoretical foundations and seminal works that have shaped the development of BRKGA, including studies on genetic algorithms, random-key representation, and elitist strategies.
The co-word analysis tracks the thematic evolution of BRKGA research over three time periods (2002-2011, 2012-2017, 2018-2023). In the early years, the focus was on genetic algorithms and fiber optic networks, while in the later periods, heuristics, computational experiments, and container loading emerged as the core and promising themes.
The review highlights the versatility and adaptability of BRKGA, which has been successfully applied to a wide range of optimization problems, including scheduling, network design, facility location, cutting and packing, clustering, vehicle routing, and machine learning. The framework's problem-agnostic approach, rapid convergence to high-quality solutions, and various strategies to maintain population diversity have contributed to its widespread adoption and continued research interest.
Stats
Genetic algorithms have demonstrated remarkable efficacy in tackling complex optimization problems.
BRKGA has been applied to several optimization problems, such as scheduling, complex network design, facility location, cutting and packing, clustering, vehicle routing, graph-based problems, and machine learning.
The primary distinguishing characteristic of BRKGA is its problem-agnostic approach, which minimizes the need for continuous redevelopment or coding of framework details.
BRKGA achieves rapid convergence to high-quality solutions through the inclusion of a double elitism mechanism within the evolutionary process.
Quotes
"BRKGA is a metaheuristic framework that uses random-key-based chromosomes with biased, uniform, and elitist mating strategies alongside a genetic algorithm."
"The primary distinguishing characteristic of BRKGA is its problem-agnostic approach. Unlike many metaheuristic algorithms where the optimization mechanism is closely tied to the problem's structure, BRKGA operates within a general framework and standard representation of solutions, which minimizes the need for continual redevelopment or coding of framework details."
"BRKGA achieves rapid convergence to high-quality solutions through the inclusion of the double elitism mechanism within the evolutionary process."