Case based Decision Making in Biologically Inspired Condition Management System

The increasing need for swift decision-making and response in condition monitoring systems is characterized by agility, scalability, and the ability to address previously unforeseen events. Two common decision-making techniques for condition management systems are procedural decision-making and deliberative decision-making. Procedural decision-making is quick and operates with minimal effort, while deliberative decision-making is slow, complex, and requires effortful attention. Case based decision making (CBDM) is a kind of procedural decision-making paradigm where new problems are solved by recalling and adapting solutions to similar past problems. It operates based on the idea that past cases can be used to inform and guide decision-making in similar situations. This research work experimentally verifies the CBDM approach for predicting future situations, utilizing analogous cases stored in memory. CBDM employs a similarity function to measure case resemblance through a novel case representation in vibration datasets. The experimental validation is conducted using a uOttawa dataset of 30 cases with time-varying rotational speed conditions. The results show promise for the initial prediction of situations based on preliminary information in a prompt manner.