A Comparison between Two Evolutionary Hyper-Heuristics for Combinatorial Optimisation

Developing and managing a general method of solving combinatorial optimisation problems reduces the need for expensive human experts when solving previously unseen variations to common optimisation problems. A hyper-heuristic provides such a method. Each hyper-heuristic has its own strengths and weaknesses and we research how these properties can be managed. We construct and compare simplified versions of two existing hyper-heuristics adaptive and grammar-based, and analyse how each handles the trade-off between computation speed and quality of the solution. We test the two hyper-heuristics on seven different problem domains using the HyFlex framework. We conclude that both hyper-heuristics successfully identify and manipulate low-level heuristics to generate "good" solutions of comparable quality, but the adaptive hyper-heuristic consistently achieves this in a shorter computational time than the grammar based hyper-heuristic.