Selected Papers from the PATAT'02, LNCS 2740. Also available in the Proceedings of PATAT'02, 90-103. Aug, 2002 Knowledge Discovery in a Hyper-Heuristic for Course Timetabling Using Case-Based Reasoning E.K. Burke 1 , B.L. MacCarthy 2 , S. Petrovic 1 , R. Qu 1 1 School of Computer Science and Information Technology, Jubilee Campus, University of Nottingham, Nottingham, NG8 1BB, U.K. {ekb, sxp, rxq}@cs.nott.ac.uk 2 School of Mechanical, Materials, Manufacturing Engineering and Management, University of Nottingham, University Park, Nottingham, NG7 2RD, U.K. bart.maccarthy@nottingham.ac.uk Abstract. This paper presents a new hyper-heuristic method using Case-Based Reasoning (CBR) for solving course timetabling problems. The term Hyper- heuristics has recently been employed to refer to “heuristics that choose heuris- tics” rather than heuristics that operate directly on given problems. One of the overriding motivations of hyper-heuristic methods is the attempt to develop techniques that can operate with greater generality than is currently possible. The basic idea behind this is that we maintain a case base of information about the most successful heuristics for a range of previous timetabling problems to predict the best heuristic for the new problem in hand using the previous knowledge. Knowledge discovery techniques are used to carry out the training on the CBR system to improve the system performance on the prediction. Initial results presented in this paper are good and we conclude by discussing the con- siderable promise for future work in this area. 1 Introduction 1.1 Case-Based Reasoning What is Case-Based Reasoning? Many techniques from Artificial Intelligence (AI) and Operational Research (OR) solve timetabling problems directly by employing heuristics, meta-heuristics and hybrids on the problem in hand [1, 2, 3]. Case-Based Reasoning (CBR) [4] is a Knowledge-Based technique that solves problems by em- ploying the knowledge and experience from previous similar cases. Solutions or prob- lem solving strategies that were used in solving earlier problems (cases) are main- tained in a store (case base) for reuse. Adaptation usually needs to be carried out for