This paper describes a new hyperheuristic algorithm that performs well over a variety of different problem classes. A novel method for switching between working on a single solution and a pool of solutions is proposed. This method is combined with an adaptive strategy that guides the selection of the underlying low-level heuristics throughout the search. The algorithm was implemented based on the HyFlex framework and was submitted as a candidate for the Cross-Domain Heuristic Search Challenge 2011.
展开▼
