Situation awareness in context-aware case-based decision support.

摘要

Humans naturally reuse recalled knowledge tosolve problems and this includes understanding the informationthat identify or characterize these problems (context), andthe situation. Context-aware case-based reasoning (CBR)applications uses the context of users to provide solutionsto problems. The combination of a context-aware CBR withgeneral domain knowledge has been shown to improve similarityassessment, solving domain specific problems and problems ofuncertain knowledge. Whilst these CBR approaches in contextawareness address problems of incomplete data and domainspecific problems, future problems that are situation-dependentcannot be anticipated due to lack of the facility to predict thestate of the environment. This paper builds on prior work topresent an approach that combines situation awareness, contextawareness, case-based reasoning, and general domain knowledgein a decision support system. In combining these concepts thearchitecture of this system provides the capability to handleuncertain knowledge and predict the state of the environment inorder to solve specific domain problems. The paper evaluates theconcepts through a trial implementation in the flow assurancecontrol domain to predict the formation of hydrate in sub-seaoil and gas pipelines. The results show a clear improvement inboth similarity assessment and problem solving prediction.