What Should Simulations Simulate?

Abstract

Both man-in-the-loop and computer simulations have become relatively mature technologies and are used quite extensively in human-machine systems analysis and design. As simulation becomes more mature and its application more routinized, the fundamental assumptions, both substantive and methodological, upon which simulations rest are sometimes neglected or forgotten. As a consequence some applications of simulation may be less than optimum and there may be some uncertainty surrounding the role which simulation can and should play in the human-machine systems domain. In this paper several types of simulations are classified according to the characteristics of the modelling performed and the representations used. These classes are briefly analyzed to establish some of the ontological and epistemological assumptions required in each class. The assumptions are then compared against the reasons and objectives usually cited for using simulation in human-machine systems work. This comparison provides useful insights into the role of simulation and enables some suggestions for needed work on modelling and theories of human-machine system performance.