Paralelismo de um algoritmo genético aplicado a otimização de rotas em mineração de modo escalável

Abstract

At present, the technological trend is focused on artificial intelligence and Internet of Things applied in embedded systems. With the constant improvement of processes that are being implemented as a business culture in most companies, the mine operating area in the mining sector has not been left out of the search for new technologies. Among the main focuses of improvements is the dispatch system, which operates in fleet management. The high cost of a route optimization system for a dispatch system still does not favor its deployment in all mining companies. This work seeks to optimize the execution time of a genetic algorithm developed specifically for the optimization of routes in open pit mines. The proposed genetic algorithm uses parallelization with the Island-type model and uses the OpenMP tool to perform the multi-process programming to be applied in small, medium and large-scale mining in the optimization of dispatch in mining with multiple routes. The management of the threads with the process of execution of the genetic algorithm requires specific data for execution, where in this proposal it has a scalable configuration according to the size of the mining company.