Improving the Performance of Queueing Systems through Cross-Training

Georgia Tech Research Corporation, Atlanta GA

Investigators

Abstract

This grant provides funding for a study of how workers should be assigned to tasks, assuming that the workers can be cross-trained to capably perform a variety of tasks. The goal is to develop easily implementable, robust, and dynamic work assignment policies that achieve improvements in performance measures such as throughput, sojourn time, and work in process. The research will also seek to identify the conditions under which particular work assignment strategies will be of greatest benefit, quantify the magnitude of the resulting performance improvements, determine the aspects of these strategies that have the greatest impact towards improved performance, and investigate the sensitivity of the results to the specific (modeling and analysis) assumptions made. In order to achieve the goals of this project, organizations will be modeled as queueing systems and analyzed using several methodologies (i.e., Markov chain analysis, Markov decision processes, stochastic comparisons, fluid relaxation, and simulation). If successful, the results of this research are expected to include detailed guidelines for how cross-training and dynamic work assignment strategies can be used to improve an organization's performance and for how much benefit can be obtained by implementing these strategies. This research is also expected to to provide qualitative insights that managers can use to implement simple changes in how their organization's workforce is utilized that are likely to yield improved performance.