Memetic Algorithms and Hyperhill-climbers

Memetic algorithms (MAs) are meta-heuristics that join genetic algorithms with hill climbing. MAs have recognized success in solving difficult search and optimization problems. Hyperheuristics are proposed as an alternative to meta-heuristics. A hyperheuristic is a mechanism that chooses a heuristic from a set of heuristics, applies it to a candidate solution, and then makes a decision for accepting or rejecting the new solution. In a traditional MA, a single hill climbing method is utilized during the search process. In the presence of multiple hill climbers, the hyperheuristic mechanisms can be adapted for the MAs and employed to exploit the strength of each hill climber better without changing the framework of the MAs. In this study, a set of such mechanisms referred to as hyperhill-climbers is investigated for solving exam timetabling problems.