9622426 Johnson The research involves developing accurate models of traffic processes in queueing models for designing and operating communications and manufacturing systems. The usual assumptions for making the models analytically tractable such as Poissonian flow is generally not adequate to capture the complex nature of real traffic processes in these applications. A methodology is needed for identifying traffic-process descriptors for gaining accurate performance measures from the queueing models. Once adequate descriptors are defined, appropriate queueing models must be developed. One such promising descriptor is traffic burstiness and research effort is required to quantify it. Then, appropriate queueing models must be developed to reflect the quantified burstiness descriptor. Finding more realistic queueing representations of traffic process in communications and manufacturing would add immeasurable to the design and operation of these crucial systems. This research, in combining process descriptor development with queueing model development, is a sound approach which should produce useful results toward this end and add to the theory of queueing models in these areas.
