Learning-enhanced adaptive DSS: a Design Science perspective

The process of acquiring, maintaining, updating, and using appropriate domain–specific knowledge has played an integral role in knowledge–based decision support systems. Although each of these stages is necessary and important, knowledge–based systems that operate in dynamic environments can become quickly stale when core knowledge embedded in these systems are not continually updated to reflect changes in the system over time. Clearly, stale knowledge could be faulty and cannot be relied upon for making decisions and a knowledge–based decision support system with stale knowledge may even be detrimental in the long run. We consider a generic adaptive DSS framework with learning capabilities that continually monitors itself for possible deficit in the knowledge–base, expired or stale knowledge already present in the knowledge–base, and availability of new knowledge from the environment. The knowledge–base is updated through incremental learning. We illustrate the generic knowledge–based adaptive DSS framework using examples from three different application areas. The framework is flexible in being able to be modified or extended to accommodate the idiosyncrasies of the application of interest. The framework considered is an example artifact that naturally satisfies the Design Science perspective.