@article{oai:ipsj.ixsq.nii.ac.jp:00059877,
 author = {Kiyoshi, Furuya and Yoshihiro, Tohma and Kiyoshi, Furuya and Yoshihiro, Tohma},
 issue = {1},
 journal = {Journal of Information Processing},
 month = {Mar},
 note = {Hakimi and Amin [2] gave necessary and sufficient conditions for identification of all faulty units in a system on the assumption that the tests are complete. We extend those theories to the case where even fault-free units may fail tests of faulty ones. We define a system to be (t r)-diagnosable if all faulty units can be identified from test outcomes  provided the number of faulty units and that of test failures do not exceed t and r respectively. Similarly  we define totally-・ﾍdiagnosability  where ・ﾀrepresents the summation of them. We first give the necessary and sufficient condition for a system  in which no two units test each other  to be totally-・ﾍdiagnosable. Then  we extend the analysis to general case  and give the condition for a system  on which no such restriction is placed  to be (t r)-diagnosable., Hakimi and Amin [2] gave necessary and sufficient conditions for identification of all faulty units in a system on the assumption that the tests are complete. We extend those theories to the case where even fault-free units may fail tests of faulty ones. We define a system to be (t,r)-diagnosable if all faulty units can be identified from test outcomes, provided the number of faulty units and that of test failures do not exceed t and r respectively. Similarly, we define totally-・ﾍdiagnosability, where ・ﾀrepresents the summation of them. We first give the necessary and sufficient condition for a system, in which no two units test each other, to be totally-・ﾍdiagnosable. Then, we extend the analysis to general case, and give the condition for a system, on which no such restriction is placed, to be (t,r)-diagnosable.},
 pages = {24--27},
 title = {On Diagnosabilities of Systems with Incomplete Test},
 volume = {8},
 year = {1985}
}