In this proposal, the PI will develop novel efficient and effective network adaptive and dynamic bandwidth allocation strategies, named as triggered strategies, for the advanced wireless networks with multiple-class calls, multiple-channel services and multiple-level quality of service (QoS) degradations. Different calls may have the different priorities for the upgrade and/or the degradation. The key idea of the proposed strategies is to give some reasonable more chances for the blocked calls in the common sense in order to get at least the minimal bandwidth requirement from one of the neighbor cells. The core theoretic concept here is developing the quasi-reversibility concept of stochastic processes with especially consideration of the wireless and mobility features of the networks. The significant result of the triggered strategies is that the equilibrium distribution of a wireless multimedia network in general can be verified to be a product form of individual cell's equilibrium distribution. The PI also plans to provide a complete theoretical analysis of the complexity for the proposed strategies and to test the actual performance of practical interest for general wireless multimedia networks with and without using of triggered strategies. The PI will compare the results under the triggered adaptive schemes with those possible results without using of the triggered strategies and will make a software package for the triggered strategies. Another idea of this proposal is to find the optimal triggered strategy in terms of the Markov Decision Process.
Intellectual merit: The proposed triggered strategies significantly improve almost all current existing bandwidth allocation strategies and importantly produce ways of finding extremely important network's performance measures, which could not be obtained in terms of the current methods in the literature. These measures, for example, may include the proportion of a class u call at cell j can only receive a level v QoS when it comes, the probability that a class u call will gain higher QoS than class v calls, degradation period ratio, degradation ratio, upgrade/degrade frequency, system degradation probability, call degradation probability and other network important measures. The proposed theoretic results will provide rigorously novel approaches for the analysis of adaptive resource allocation strategies for wireless multimedia networks which are in great demand.
Broader impacts: Our research results will be disseminated in leading journals and conferences, and will be incorporated into courses and seminars. Via the training of students, and via collaborations and technical interactions with colleagues, both in the US and abroad, the work will lead to yet further technical progress. There are ongoing collaborations with colleagues in Canada, Japan, and China. This project will lead to the incorporation of various novel technical tools into independent studies, courses, and research seminars, which will be part of the training of many students in the department of electrical engineering and computer science. For example, this project will directly impact the capabilities, interests, and careers of the graduate students who participate in the project. At the University of Toledo, a new Ph.D. program in Computer Science and Engineering is in the final stages of approval and implementation. The success of this project would give the program a very strong boost.
