Desenvolvimento de um modelo de algoritmo genético para otimização do custo na fase de transporte logístico do pequeno empreendedor

Abstract

Genetic algorithms are metaheuristic search algorithms that simulate the evolutionary process described in Charles Darwin's theory of evolution. According to the theory, individuals with better adaptations to the environment are the ones that survive. Thus, this idea was applied to the computing scenario by John H. Holland, who in 1975 created a computational model based on biological evolution to solve optimization problems. The simplicity of his algorithm allowed for approximate solutions to NP (non-polynomial) problems to be found, attempting to find an optimal solution, but without guarantees that the solution found will be the best. The search for the problem solution is constructed through a series of functions that he defined as genetic operators: population creation, selection of the fittest, combination, mutation, and substitution. At each process, a descendant population is created that could be the potential solution, but it will go through the objective function evaluation stage until a satisfactory solution is found. Due to its flexibility and problem-solving ability, genetic algorithms can be used in various fields of knowledge, including Game Theory, Artificial Intelligence, finance, Data Science, and Engineering, optimizing processes, models, and learning. In Brazil, microentrepreneurs and small businesses face significant challenges in the logistics sector, reflecting the lack of government investment, low capital for strategic planning, and deficient infrastructure. This results in increased operational costs, particularly in the transportation phase of goods. Thus, this work aims to apply the concept of genetic algorithms to the development of an evolutionary computational model that assists in reducing the logistics operational costs of small entrepreneurs in the transportation phase of goods.