Ant-Q hyper-heuristic approach for solving 2-dimensional Cutting Stock Problem

Hyper-heuristics are new approaches which aim at raising the level of abstraction when solving combinatorial optimisation problems. In this paper we introduce a new hyper-heuristic model, namely Ant-Q hyper-heuristic, which transliterates the significant learning ability of Ant-Q algorithm proposed by Gambardella and Dorigo, for building good sequences of low-level heuristics aimed at gradually constructing final solutions. This approach was applied to 2-dimensional Cutting Stock Problem and tested through a large set of benchmark problems. The results have shown that the Ant-Q hyper heuristic is able to outperform single heuristics, well known metaheuristics and be competitive to other hyper-heuristics from the literature.