Evolutionary timetabling using biased genetic operators

Evolutionary algorithms have proved to be a flexible and effective technique for addressing various instances of the timetabling problem. We describe an investigation and its result on an evolutionary approach to solving a particular class of highly constrained timetabling problems. The convergence speed of the evolution program has been significantly improved with the usage of biased operators, which generate offspring by preserving the building blocks of the parents. We also describe two metrics for measuring the efficiency of the genetic operators and how convergence speed has been improved by applying these metrics to fine-tune the probability of the genetic operators. Computational experiments over real test problems showed promising results.