A stochastic hyper-heuristic for optimising through comparisons
Kieran Greer · 2010
This paper introduces a new hyper-heuristic framework for automatically searching and changing potential solutions to a particular problem. The solutions and the problem datasets are placed into a grid and then a game is played to try and optimise the total cost over the whole grid, using a randomising process. The randomisation could be compared to a simulated annealing approach, where the aim is to improve the solution space as a whole, possibly at the expense of certain better solutions. It is hoped that this will give the solution search an appropriate level of robustness to allow it to avoid local optima.