Methods of Predicting Human Reliability in Man–Machine Systems

Abstract

Four studies describing the development and application of a simple multiplicative probability model for human error prediction are reviewed and evaluated. The methodology is an elementaristic one that requires analysis of system operations to the task-element level. Estimates of human performance reliability are applied through the use of the Data Store, which is based on extrapolation of results from 164 experimental studies. The various possibilities of operator action are explored in detail through a “probability” tree; each branch of the tree represents an alternative contingency. Performance reliabilities for task elements are progressively combined through the use of the series product rule to yield reliability estimates for tasks, mission phases, and the overall system. The methodology has been applied in two contexts: the generation of reliability estimates through a computerized Monte Carlo program and by using experts' rating of operator performance.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 47 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.