Optimal State Detection Policies for Coherent Systems.
Abstract
The problem of minimizing the expected cost of identifying the state of a coherent system (as functioning or failed) is considered. The system is composed of components, and only individual components can be tested. Efficient algorithms are presented for some special cases of coherent systems: series, parallel, parallel-series, series-parallel and k-out-of-n systems. The concept of the Importance of Components is used to develop a branch and bound algorithm which determines the optimal testing procedure for any general coherent system. However this algorithm is not always efficient in solving this problem. Some other closely related problems are discussed, such as how to identify the state of a system which is represented as a fault tree, or how to identify the failed components when the system is known to be failed.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 1977
- Accession Number
- ADA078824
Entities
People
- Yosi Ben-dov
Organizations
- University of California, Berkeley