Constructive and local-search based hyperheuristics: A case for hybridisation?

Hyperheuristics can be defined to be heuristics which choose between heuristics in order to solve a given optimisation problem or class of optimisation problems[5, 1]. One aim of using hyperheuristic methods is to achieve robustness, that is, to generate good-quality solutions for various problems or problem instances using the same method with very limited problem-specific knowledge. Over the past half decade or so, a number of hyperheuristic methods have been developed and applied successfully to a number of relevant combinatorial optimisation problems including timetabling, scheduling, bin packing, rostering, and space allocation [5, 1, 2, 7, 6, 4]. This development has gained increasing recognition in the operational research and artificial intelligence communities (See [10, 1] for an overview of the literature on hyperheuristics). Of interest to us are two types of hyperheuristics which were recently developed at Napier and Nottingham universities: