Adaptively Parameterised Hyperheuristics for Sales Summit Scheduling

The concept of a hyperheuristic was recently proposed by the authors as an approach that operates at a higher level of abstraction than current metaheuristic approaches. The hyperheuristic chooses which low-level heuristic to apply depending upon the characteristics of the part of the search space being explored and the performance history of each low-level heuristic. In this paper we present various hyperheuristic and heuristic methods and demonstrate their effectiveness by way of a real-world case-study problem of scheduling. In particular, we describe a self-adaptive hyperheuristic which uses a choice function to rank the low-level heuristics. Results produced by the choice function hyperheuristic appear to be not only superior to those produced by other heuristic methods, but also of much higher quality than those obtained from the greedy heuristic currently used by the problem owner.