Comparison of Evolutionary and Rule-Based Strategies for Electromagnetic Device Optimization

The use of optimization techniques to improve the design of electromagnetic (EM) devices is becoming very common. However, choosing an appropriate strategy to carry out an optimization process within a practical design activity is still a pending problem. This paper proposes a novel knowledge-based strategy and compares its performance against a standard evolutionary approach. The approaches are implemented as part of a case-based design system and are tested on an induction motor design.