Evolutionary timetabling with a priority evolutionary timetabling with a priority-based indirect representation based indirect representation

Abstract

When applying evolutionary algorithms to optimization problems, the solution representation issue is of critical importance to the overall performance. This work presents an indirect representation method for genetically encoding partially described solutions of educational timetabling problems that are completed by a “timetable builder” algorithm using priorities to allocate events. A genetic algorithm using this technique is applied on several real problem instances retrieved from Greek universities, and the solutions are compared to the corresponding manmade ones, with promising results.