Sequential Test Strategies with Unreliable Tests

With the increased recognition of importance of design for testability, there is an increasing trend toward the development of efficient test strategies for system maintenance. An important issue in practical situation is the imperfect nature of tests, i.e., the tests can have missed detections and false alarms. An error diagnosis may result in false removal and missed repair of the components, thus, increases the cost of maintenance. The previous works have been studied on the sequential test strategy problem but are commonly lack of scientific evaluation system for diagnostic strategies. Furthermore, the expected testing cost is treated as the unique criterion to evaluate test strategies, but the diagnostic accuracy is another important criterion when considering test uncertainty (unreliable tests). The paper mainly considers the mentioned problems, and presents several evaluation functions for sequential test strategies, introduces two types of diagnostic strategies optimization generation methods with unreliable tests, emphasized particularly on test cost and diagnostic accuracy respectively. The simulation computational results validate the reasonability and practicability of the division. The proposed strategies can be applied to guide the option of the diagnostic strategies for maintenance engineers scientifically.