University of Manitoba

Published purpose

The design of a high-performance micro-electromechanical system (MEMS) requires extensive design simulation to optimize its performance. Complex MEMS implementations, such as in adaptive optics (AO) systems, have multiple competing specifications and structural parameters to optimize. Skilled designers often spend many months of effort to develop such MEMS systems, with design effort often stopping when performance is sufficient as opposed to fully optimized. This slows and limits innovation in applied MEMS technology areas. This Project will focus on the development of AI using multi-objective optimization to automate the design and optimization of complex MEMS systems. The multi-objective MEMS performance problem is one where AI will provide dramatic benefits and significantly accelerate innovation.