A tabu search approach for graph-structured case retrieval

In case-based reasoning (CBR), graph-structured representations are desirable for complex application domains such as planning and design. Graph is a powerful data structure and allows knowledge to be encoded completely and expressively. However, the advantages come with a computational overhead for case retrieval, which presently prevents the usage of graph-structured representation for large-scale problems. In this paper, we describe a two-stage strategy that is based on the tabu search for solving the graph-structured case retrieval. The experiments were performed on synthetically generated graphs. The preliminary results obtained show the effectiveness of our approach.