Improving Complex Adaptations in Process-Oriented Case-Based Reasoning by Applying Rule-Based Adaptation

Adaptation is a complex and error-prone task in Case-Based Reasoning (CBR), including the adaptation knowledge acquisition and modeling efforts required for performing adaptations. This is also evident for the subfield of Process-Oriented Case-Based Reasoning (POCBR) in which cases represent procedural experiential knowledge, making creation and maintaining adaptation knowledge even for domain experts exceedingly challenging. Current adaptation methods in POCBR address the adaptation knowledge bottleneck by learning adaptation knowledge based on cases in the case base. However, these approaches are based on proprietary representation formats, resulting in low usability and maintainability. Therefore, we present an approach of using adaptation rules and rule engines for complex adaptations in POCBR in this paper. The results of an experimental evaluation indicate that the rule-based adaptation approach leads to significantly better results during runtime than an already available POCBR adaptation method.