The traditional system design methodology (which follows a preliminary design, analysis, appraisal and redesign pattern) for automotive systems makes balancing contradicting design constraints such as cost, reliability and performance very difficult. Often all of the constraints are not met, and, even if they are, the resultant design may be sub-optimal.
This paper outlines a new design approach using constraint satisfaction methods to create an optimal design, where all of the constraints are satisfied. Once the constraints are identified and represented in a suitable format, the design task can be semi-automated using a suitable algorithm.
This paper explains how to identify and represent constraints in an automotive problem, the algorithms that can be used to solve the design problem, and how the outcome can be used as an extension to the existing design methodology. Finally, a high-level design for the implementation of the automation process and details of an initial prototype are given.