Incorporating chaos into the developmental approach for solving the examination timetabling problem

Abstract

The developmental approach DA is a novel methodology which mimics processes from cell biology. Previous work has evaluated different versions of the DA to solve the examination timetabling problem. The DA creates an organism, which represents a solution to the problem, using the processes of cell division, cell interaction and cell migration. A shortcoming of the standard DA algorithm is that is does not incorporate any form of chaos or disorder which is essential for organism development in nature. The paper firstly presents and tests noise operators for this purpose. It then examines the use of an evolutionary algorithm to evolve noise in the DA. The performance of the standard DA, the DA using noise operators DANO and the DA evolving noise DAEN are tested on a set of ten real-world examination timetabling problems from the Carter benchmark set of problems. DAEN was found to outperform both the DA and DANO. Furthermore, the results produced by DAEN were found to be comparative to other methodologies applied to the same set of problems.