Measuring Metaheuristic Performance over Timetabling Problem Instances Using Fitness Distance Correlation Method

The Timetabling Problem is a combinatorial optimization problem. Over the last decade variant of Metaheuristic approaches have been used to solve various type of Timetabling Problem with great success. Metaheuristic are approximate algorithm used when the size of the search and spaces becomes unmanageable. However the stochastic nature of metaheuristic make it difficult to measure the performance of the algorithms. In this paper we analyze the performance of Hybrid Evolutionary Metaheuristic for a set of benchmark timetabling instances through an analysis of the fitness-distance correlation (FDC). FDC is a statistical measure of a search difficulty in relation to Genetic Algorithm. We propose new ways of FDC analysis. Initial result indicates that FDC can also be used in different ways of analysis as well as different problem instances. Key words