Uso do algoritmo genético e recozimento simulado para o problema de alocação de salas

Abstract

This work address the implementation and comparison of two heuristic techniques for Class Assign Problem: Genetic Algorithm and Simulated Annealing. Since it is a NP-hard problem, some heuristic methods have been proposed to solve it. The heuristic methods search for good solutions in reasonable time, when exact methods are appropriated. The Class Assign Problem considered involves additional restrictions when compared to others previous similar works. These restrictions are the covered distance by students to walk from a room to another and the allocation of practical lessons. The best results for each method are compared observing same set of local search operators. To do end, some conclusions are possible and suggestions are presented for possible improvements on both algorithms.