Development and application of hyperheuristics to personnel scheduling
This thesis is concerned with the investigation of hyperheuristic techniques. Hyperheuristics are heuristics which choose heuristics in order to solve a given optimisation problem. In this thesis we investigate and develop a number of hyperheuristic techniques including a hyperheuristic which uses a choice function in order to select which low-level heuristic to apply at each decision point. We demonstrate the effectiveness of our hyperheuristics by means of three personnel scheduling problems taken from the real world. For each application problem, we apply our hyperheuristics to several instances and compare our results with those of other heuristic methods. For all problems, the choice function hyperheuristic appears to be superior to other hyperheuristics considered. It also produces results competitive with those obtained using other sophisticated means. It is hoped that - hyperheuristics can produce solutions of good quality, often competitive with those of modern heuristic techniques, within a short amount of implementation and development time, using only simple and easy-to-implement low-level heuristics. - hyperheuristics are easily re-usable methods as opposed to some metaheuristic methods which tend to use extensive problem-specific information in order to arrive at good solutions. These two latter points constitute the main contributions of this thesis.