Abstract
The modern concept of New Product Development (NPD) cycle is often described as a funnel with continuous local divergence and convergence stages. Despite the ongoing research on Cased Based Reasoning systems and Expert Systems in NPD processes, the utilization of decision support tools, especially in the part selection stage, is rather minimal. Hence, a new macroscopic bearing selection system is proposed to illustrate the feasibility of reinforcing two issues with the current part selection process and Knowledge Based Systems (KBS): a lack of part selection decision support tool for beginner designers for whom con?dent part selection at the early stage of prototyping proves challenging, and the current KBS systems’ inability to transfer their expertise to the human designers. The proposed part selection system achieves this in two stages: the ?rst stage being the rule knowledge base that narrows down the ?eld of interest and de?nes the new problem according to the chosen attributes, which is followed by the second stage of presenting the knowledge base in user digestible forms. A decision tree is developed to demonstrate the feasibility of implementing the rule knowledge base for the ?rst stage, and illustrations of the second stage are showcased for the future development. As a result of these two stages, the proposed macroscopic part selection system seeks to achieve a di?erent goal from other con-ventional knowledge based systems. The goal of the proposed system is not to make the correct decision, but to expedite the learning process of the designer by providing focused and organized information that pertains to the selection group to which the designer’s product belongs to. Hence it aims to assist the designer’s decision-making logic, rather than the making of the decision itself so that the system escapes from be-ing a ‘black box’ with a simple input and an output, and becomes more of a ‘formula book’ that teaches the beginner designers the process of part selection.