Multi-case-base reasoning

This dissertation explores a novel distributed case-based reasoning architecture we term multi-case-base reasoning (MCBR). Case-based reasoning (CBR) systems solve new problems by retrieving stored prior cases, and adapting their solutions to fit new circumstances. Traditionally, CBR systems draw their cases from a single local case-base tailored to their task. However, when a system's own set of cases is limited, it may be beneficial to supplement the local case-base with cases drawn from external case-bases for related tasks. Further, as deployed case-based reasoning systems become increasingly prevalent, opportunities will arise for supplementing local case bases on demand, by drawing on the case bases of other CBR systems. Effective use of external case-bases requires strategies for multi-case-base reasoning : (1) for deciding when to dispatch problems to an external case-base, and (2) for performing cross-case-base adaptation to compensate for differences in the tasks and environments that each case-base reflects. This research establishes the component processes required for MCBR, and contrasts MCBR with traditional CBR. We present knowledge-light algorithms for automatically tuning MCBR systems by selecting effective dispatching criteria and cross-case-base adaptation strategies. We present experimental illustrations of the performance of the tuning methods for a numerical prediction task, and demonstrate that a small sample set can be sufficient to make high-quality choices of dispatching and cross-case-base adaptation strategies. To examine the relative tradeoffs between MCBR and merging the individual case-bases into a single case-base, we provide an experimental assessment of how case-base merging may be detrimental and how MCBR affects the quality of solutions generated. This demonstrates that for a given local case-base and an external case-base for a task environment that is similar to, but different from, the local task environment, MCBR can improve accuracy compared to merging the case-bases into a single case-base.