Optimization of Call Centre Scheduling using the Cross-Entropy Method.

Abstract: This thesis project investigates how the cross-entropy method can be applied to the optimization of scheduling problems. It is an important question because lower staffing leads to lower operating cost. Scheduling problems cannot be solved exactly in reasonable time, and finding methods to solve them efficiently is of great importance. The cross-entropy method has been used with good results for other problems with time-consuming properties. The idea behind the cross-entropy method is to generate a random sample with possible solutions, and calculate how well they performed. Then that information is used to generate a new random sample with greater probabilities around good solutions. The method provides a formula how to update the probability distributions from which the random samples are generated. The method can be applied to create a schedule for a call centre. However, it is uncertain how good the solutions are, what complexity the method has, and how differences in the problem specification will affect the method’s performance.