Modeling the Impact of Workload in Network Centric Supervisory Control Settings

Abstract

The Department of Defense’s vision of network centric operations will likely bring about higher operator mental workload due to the large volume of incoming information. As a result, it is critical that systems designers develop predictive models of both human and system performance, such that they can determine how a proposed technology will influence not only operator workload, but also system performance. To this end, this paper introduces a discrete event simulation approach to human-system modeling that includes a quantitative relationship between workload and performance, inspired by the Yerkes-Dodson relationship. Using the concept of utilization, or operator percent busy time, as a surrogate workload measure, we demonstrate that a quantitative instantiation of an inverted-U workloadperformance curve improves discrete event simulation model predictions in a human supervisory control model of single operator control of multiple unmanned vehicles. While this work generates the first known empirical evidence of a parabolic workload-performance as a variable in a quantitative predictive human performance model, this effort is preliminary and future research implications are discussed.