Optimising a presentation timetable using evolutionary algorithms

Abstract

This paper describes a solution to the problem of scheduling student presentations which uses evolutionary algorithms. The solution uses a permutation based approach with each candidate schedule being coded for by a genotype containing six chromosomes. Five systems (chromosome representation and genetic operators) are described and their suitability assessed for this application. Three of the systems use direct representations of permutations, the other two use indirect representations. Experimental results with different fitness equations, operator rates and population and replacement strategies are also given. All the systems are shown to be good at solving the problem if the algorithm parameters are correct. The best parameters for each system are given along with those parameters that do particularly badly.