Robert G. Feyen

A Conceptual Framework for Dynamic Prioritization in Multiple-Task Scenarios

Prioritizing tasks appropriately is particularly critical when performing multiple tasks concurrently. Although necessary to achieve ones goals or avoid serious consequences, prioritization has not received much attention in the research literature, especially with respect to modeling human performance computationally. A conceptual framework that integrates several motivational theories, empirical studies, and neuroscience research is proposed to guide future studies of dynamic prioritization in multiple-goal contexts. Rooted in control theory, the proposed framework illustrates self-regulation processes in prioritizing tasks and explicitly shows important factors affecting the prioritization process so that empirical results can be integrated into the framework and future studies can be inferred. By illustrating information flow in the self-regulation processes and the brain structures associated with prioritization, the framework should help facilitate development of robust computational models of task prioritization.

A queueing network model of task prioritization using a general hierarchy of prioritization rules

Earlier, Zhang and Feyen (2005) proposed a qualitative framework for predicting how people working in a multitasking scenario switch between concurrent tasks with dynamically changing priorities. This paper describes a validation study of a computational model derived from this framework. Utilizing a general hierarchy of prioritization rules suggested by a companion empirical study, a model of a multiple task scenario built using a queueing network approach was compared to the empirical results. On all metrics considered, no means were found significantly different and the model replicated all but one of 54 task sequences demonstrated by human subjects. Comparisons to similar models utilizing only single prioritization rules revealed that the general hierarchy yielded substantially better predictions.