Construção e implantação de um sistema automático de alocação de espaço físico baseado em computação evolutiva para a Universidade Federal de Uberlândia

Abstract

Accomplished twice per school year, the rooms allocation process at the Federal University of Uberlândia (UFU) is an activity with a high degree of complexity. The activity consists of the distribution of a large number of courses into available classrooms for allocation. The administrator of institution physical space must attend a number of constraints defined by the direction and, as far as possible, the needs of each course. The constraints make the entire process longer, taking months to be done. The present work describes the space allocation problem (SAP) and the existing particularities in the process of UFU and other universities, it was verified by comparing SAP in different institutions that they handle different objectives and, therefore, they can’t be solved using the same approach. A suitable search method for complex optimization problems as SAP, are genetic algorithms (GA). Analyzing the existing ways for handling constraints in GA along with the considered objectives for space allocation in UFU, it was decided to use a special representation of the individual and also specific genetic operators for the problem. A system for automating the SAP in UFU was built and deployed, the obtained results using university’s real data indicated the feasibility of the method, besides reducing the time required for carry out the process.