| Popis: |
Consider a network G of n units, each of which can be tested by those units from which there is a test connection. The outcome of each test is binary (good or faulty) and is the judgment of the testing unit on the tested unit. We present an algorithm for identifying the minimum number of faulty units based on the test outcomes. It works in time proportional to τ(G)m, provided the number of faulty units is no more than τ(G), where m is the number of test connections and τ(G) is a parameter of G such that if the number of faulty units is no more than τ(G), then they are uniquely identifiable. |
