SAUL BLUMENTHAL

I work in the area of reliability and failure distributions. Suppose that a transistor, for example, is either defective or perfect and the only way to tell is to test it until it fails. If we test all the transistors in a fixed batch for a specified time period, we can observe the times at which some items fail. Based on this data, I have studied ways to estimate the failure distribution and the number of defectives among the items that have not yet failed. In Blumenthal and Dahiya (1995), we deal with the problem where not all of the failure times have been recorded. An unrelated problem that I studied in Blumenthal (1984) has the following application: choosing the manufacturing process for producing a certain type of transistor that gives the longest average lifetime and estimating the value of this average. In Zhou and Blumenthal (1992), we have studied the use of new systems to test what the reliability will be for older, mature systems (the problem being that the new systems tend to look more reliable than they will be after they have aged).