May 2007 to September 2010
Funded by a leading blue chip telecommunications company, this project at the School of MACS, Heriot-Watt University, is to investigate new ideas in the area of `hyper-heuristics' and apply these to the task of discovering new fast algorithms for a specific range of data mining applications of relevance to the sponsors. Hyper-heuristics research originated in the UK as a by-product of research at Edinburgh involving Hsia-Lan Fang, David Corne, and Peter Ross, and uses (usually) population-based meta-heuristic search to discover algorithms that solve a specific family of problems. In that early work the idea was applied in logistics problems such as timetabling and scheduling. This work has received much take-up in the academic and industrial communities, with hyper-heuristic based solutions in logistics turning out to be eminently practical and successful. In the current `blue skies' project, the idea is to see how these notions can be used in data mining. We will be developing and investigating various ideas, guided in part by strategic requirements and specific data sources emerging from our research sponsors.
Project Team:
Principal Investigator: David Corne; Research student: Alan Vella