Design of Intelligent Adaptive Network State Updating Mechanism for the Optimization of QoS Network Performance Under Uncertainty

摘要

Large-scale communication networks are complex dynamical fuzzy systems. To architect these networks as intelligent systems, intelligence must be incorporated into their individual communicating components. The sought intelligence can be applied to the modeling and design, the maintenance of state information, and the routing protocols. In a large-scale dynamical network, the global network state information is inherently imprecise. The nature and the level of the inevitable imprecision vary significantly when different information updating policies are applied, which directly affects the routing decisions and the network performance. Nevertheless, the trend in QoS Networking research has been to focus on developing routing algorithms while regarding the updating policies as supplements to the routing algorithms to enable their implementation. In this paper, we reverse the tradition by setting the primary goal to be the design of an intelligent and adaptive network state updating policy. The proposed mechanism, regardless of the routing algorithm, aims at the optimization of the network performance by dynamically maintaining the optimal trade-off between the degree of imprecision and the overhead associated with the updates. The proposed mechanism is based on the use of fuzzy overlapping tolerance classes and is designed to implement bounded uncertainty and controlled overhead without the need for hold-down timers.