Optimal Admission and Pricing in Non-stationary Queueing Systems

点击翻译按钮获取中文摘要

This grant provides funding to advance the state of the art in optimal control of non-stationary systems. Specifically, the objective of this research is to develop optimal admission and optimal pricing policiesfor finite capacity, non-stationary queueing systems with multiple customer classes offering different amounts of reward to obtain service. These types of queueing models arise in many applications such as production systems, communications systems, and airline industry. The research will involve thefollowing tasks. First, systems where parameters such as rewards, system capacity, arrival rates, and service rates change only once at a known time, will be considered. Insights gained from the structure of optimal policies in these simple non-stationary systems will facilitate the analysis of the systems where changes occur periodically. Finally, numerical algorithms will be developed to compute the parametersof optimal policies. A major contribution anticipated from this research is the development of a fundamental science base required for optimal decision making in non-stationary environments. Towards this end, new optimality criteria for non-stationary systems will be defined. The educational impact of the research will be in the form of classroom instruction, PhD student research, seminars and conferences.