Output details
11 - Computer Science and Informatics
University of Stirling
On the Automatic Discovery of Variants of the NEH Procedure for Flowshop Scheduling Using Genetic Programming
<22> This paper presents an automatic genetic programming approach to design specialised variants of the most successful constructive heuristic for the well studied flow shop scheduling problem. The proposed methodology significantly outperforms the original heuristic on the benchmarks studied. Once a variant of the heuristic, targeted to a class of instances, is discovered, it can be applied to quickly solve a new instance. The exploration of genetic programming to generate new heuristics plays a key role in a new major EPSRC programme grant (EP/J017515/1) of £6.8M between UCL, Stirling, York and Birmingham, which started in 2012.