Catalogo Articoli (Spogli Riviste)

OPAC HELP

Titolo:
SIGNAL-DETECTION FOR HUMAN ERROR-CORRECTION IN QUALITY-CONTROL
Autore:
SYLLA C; DRURY CG;
Indirizzi:
NEW JERSEY INST TECHNOL,SCH IND MANAGEMENT NEWARK NJ 07102 SUNY BUFFALO,DEPT IND ENGN BUFFALO NY 14260
Titolo Testata:
Computers in industry
fascicolo: 2, volume: 26, anno: 1995,
pagine: 147 - 159
SICI:
0166-3615(1995)26:2<147:SFHEIQ>2.0.ZU;2-Y
Fonte:
ISI
Lingua:
ENG
Soggetto:
SAMPLING PLANS; INSPECTION; VIGILANCE;
Keywords:
QUALITY CONTROL; SIGNAL DETECTION THEORY; HUMAN FACTORS; HUMAN ERROR COMPENSATION; INDUSTRIAL PRODUCTION SYSTEMS;
Tipo documento:
Article
Natura:
Periodico
Settore Disciplinare:
CompuMath Citation Index
Science Citation Index Expanded
Citazioni:
43
Recensione:
Indirizzi per estratti:
Citazione:
C. Sylla e C.G. Drury, "SIGNAL-DETECTION FOR HUMAN ERROR-CORRECTION IN QUALITY-CONTROL", Computers in industry, 26(2), 1995, pp. 147-159

Abstract

Acceptance sampling by attributes is now a universal and fundamental tool in statistical quality control. Among the various plans, single sampling has been recognized to offer significant cost savings and administrative simplicity. Single sampling by attributes is generally carried out or supervised by a human inspector. Unfortunately, human inspection error is another fact of life in industrial inspection. Two types of errors are possible. One is the Type 1 error, in which a good item is classified as defective (i.e., nonconforming); the other is the Type 2 error, in which a defective item is passed as good. Presently, little has been written on the effects of inspection errors on the outgoing quality and on how to compensate for these errors. The research objective of this paper is to show how a knowledge of the errors can beused to design compensating acceptance sampling plans for industrial quality control tasks. In contrast to the typical human inspection models in quality control, with fixed performance irrespective of incoming product quality, the model proposed here uses a form of signal detection theory (SDT) to predict inspector performance in order to improvesystem performance. The paper proposes the concept of lability, measured by a number ranging from 0 to 1, to characterize an inspector's ability to response to costs, rewards and probabilities involved in the inspection decision. This concept allows the modeling of both a constant-error inspector having a lability of 0, an SDT inspector having a lability of 1, and all those which fall within this range. Using this concept, the paper presents two models of the human inspector with considerations of: (a) constant Type 1 and Type 2 errors independent of the incoming fraction nonconforming, and (b) increasing Type 2 error anddecreasing Type 1 error with increasing fraction nonconforming as specified by SDT. A general model covering these inspection errors is proposed. The paper shows that an imperfect inspector, who can detect thesignal properly to adjust the decision criterion conservatively in the direction specified by the SDT, can compensate for the potential errors. A numerical example is proposed to illustrate the model application.

ASDD Area Sistemi Dipartimentali e Documentali, Università di Bologna, Catalogo delle riviste ed altri periodici
Documento generato il 15/07/20 alle ore 15:00:51