2006 Published by Elsevier Inc. defective coins, it is immensely more difficult to search the optimal testing algorithm for m defective coins.

www.elsevier.com/locate/amc0096-3003/ $- see front matter 2006 Published by Elsevier Inc.Cairns in [5] and Tos̆ic ́ in [6], studied the group testing problem of identifying two uniform defective coins, and proposed a testing algorithm respectively. But neither of them related to the group testing problem of identifying or selecting two different defective coins. Hwang only studied the basic model of identifying two different defective coins and proposed a testing algorithm in [3], while leaving the group testing problem of the other models of identifying two different defective coins open. In this paper, the authors studied the group testing problem of identifying two different defective coins in a set of k coins. A testing (weighing) algorithm for all the possible values of k would be constructed, and here kP 3, k 2 N. Let IkDCð2ÞðpÞ denote the minimum number of testing of identifying two different defective coins and SkDCð2ÞðpÞ denote the minimum number ofKeywords: Defective coin; Standard coin; Group testing; Information-theoretic lower bound; Information feedback 1