Optimization of yield and reliability using enhanced worst-case methods

This book presents a novel framework on analysis and optimization of Yield & Reliability in microsystems. The key contributions of this work are the enhancement of the worst-case method to enable its usage in yield analysis of microsystems, and a methodology that could be used as a best practice method to optimize yield, and the introduction of a new coefficient namely, the Worst-Case Reliability Coefficient (WCRC) that quantifies reliability of a microsystem relational to the design space of its operational parameters. The introduction of WCRC is crucial because this enables the designers to compare the reliability of different designs, thereby encouraging and empowering Design for Reliability approaches. The vital advances made to the worst-case method have been implemented into a comprehensive framework named „Worst-Case Analysis Suite“ (WCAS). This framework enables analysis of systems with linearized and non-linear based specification boundaries, thereby, allowing a comparison in results of the pre and post enhancement scenarios. WCAS makes use of a powerful set of optimization algorithms like the Genetic Algorithms and the Direct Search Algorithms. Hence robust optimization could be assured while solving constrained non-linear multivariable optimization problems in the worst-case analysis. WCAS has been used to demonstrate an improvement in accuracy of the enhanced worst-case methods in comparison to the established. Also, with the implementation of the enhanced worst-case method, the advantage of not having to wait long times to estimate a trustworthy yield number was exploited by the WCAS framework. WCAS has been successfully deployed in the optimization of design yield for multiple microsystems like thermal actuators, RF switches, capacitive sensors, etc.