Modeling of man-machine systems. A challenge for systems analysis

Recent industrial accidents have demonstrated the need for considering also human acts influencing the safety of industrial installations. An understanding of human errors will rely on an understanding of the cognitive mechanisms of the human decision maker. Different attempts have been made to build models of the human operator, but the models have mostly concentrated on single individuals and isolated tasks. The models developed are also rather descriptive and are not very useful in guiding the design of new systems. At present there are no models by which accurate predictions of the performance of a socio-technical system can be given. It is necessary to develope new modeling techniques, before it will be possible to model the human parts of the industrial installations to such a degree of accuracy that it is possible to give predictions on their safety. The paper discusses different modeling approaches and their shortcomings in providing guidance for the design of future systems. The paper also discusses some general requirements on the modeling methods of man-machine systems