Uma proposta de solução em hardware para o problema das p-medianas utilizando algoritmo genético e unidades de processamento softcore

Abstract

Decision making is a recurrent activity in the daily commercial, military and industrial, occurring in different instances and facing different problems. Operational research tools are proposed in the literature to assist in the search for better results in decision-making processes. From these tools, it is expected that the response time of these tools will meet criteria of feasibility of application. In the problematic context of facility allocation, the modeling of p-medians and the genetic algorithm optimization tool stand out in the development of computational solutions for complex mathematical problems. In the context of development of computational solutions, the form of parallel computing stands out for the best performance against sequential computation of algorithms. Among the ways to implement parallel computing, the use of reconfigurable hardware platforms stands out for flexibility, low cost and high performance, achieving significant reductions in processing time. However, until then, the application of parallel computing architecture in reconfigurable hardware to solve the problem of p-medians using the genetic algorithm was not proposed in the literature. In this work, we propose a reconfigurable hardware computing system, based on the Nios II softcore processor, to implement a solution to the p-median problem using the genetic algorithm. . To evaluate the proposed system, we use the metrics of accuracy and processing time.