Summary of Recent Research Accomplishments on 'Stochastic Network Processes'

摘要

The aim of this research has been to develop stochastic network processes formodeling the movement of discrete units in networks. Primary examples are the movement of data packets in computer networks, the movement of parts and supplies in manufacturing plants or in military support systems, and the movement of smart cars and trucks on electronically monitored highways. The distinguishing feature of our research is the emphasis on the next generation of intelligent networks that will be the backbone of our computer, military and transportation systems. Most of the present theory of stochastic network processes is for unintelligent networks in which the nodes operate independently, the routes of units are independent and the units move one at a time. In an intelligent network, however, the processing at the nodes and the routing typically depend dynamically on the actual congestion, and units move concurrently. Examples of dependencies are routing units to avoid congested nodes, speeding up of processing as queues grow, splitting and merging of units, batch processing and distributed as parallel processing. Our general goal is to provide an understanding of intelligent networks by describing their stochastic behavior. (kr)