Fuzzy optimization models for analog test decisions

Test decisions still constitute one of the most difficult and time-consuming design tasks. This is particularly true in the analog domain where some basic test questions have not yet been completely resolved. Since the gap between a good and a bad analog circuit is not always well-defined, extensive tests may result in the rejection of many fault-free ICs. The objective of this article is to propose fuzzy optimization models that can help in the more realistic formulation and resolution of the analog test problem. The set of good or fault-free ICs is considered as a fuzzy set. Each performance test is represented by a membership function. A global test measure is obtained by aggregating all the performance tests. An illustrative example using these concepts is provided.