Research On End-to-End Resource Allocation Technology Of Cognitive Network

The future of the communications network will move towards the direction of intelligence and adaptability,cognitive network in this context came into being.Cognitive network is a kind of intelligent network system,it can be aware of the current network conditions and the future changes in the network to predict the situation,and on this basis to make more intelligent decision-making,improve resource utilization.But at the same time,the cognitive network of network resource management allocation method also put forward new requirements.The paper mainly studies the end-to-end resource management and distribution technology of cognitive network.The main work of this paper includes the following aspects:(1)In the cognitive network,can affect the quality of business services,a variety of network factors should be regarded as the resources that the cognitive network can control.In this paper,the analysis of end-to-end transmission scenarios,the main selection of the most important link resources and node resources as the object of control.Through the characterization of link resources and node resources,the definition of resources is extended by traditional networks.And then link resources and node resources associated with the treatment,a unified management of the distribution basis.Finally,the resource allocation algorithm model is proposed according to the characteristics of cognitive network.(2)Research on end-to-end traffic prediction algorithm.In this paper,the Markov flow forecasting algorithm is proposed to improve the Markov flow forecasting algorithm.Firstly,the end-to-end flow samples are clustered by fuzzy clustering,and the membership matrix is obtained.Then the Markov transition probability matrix is established according to the membership degree matrix,and the end-to-end traffic prediction is obtained according to the established transfer model.(3)The resource allocation strategy of cognitive network is discussed in detail.Aiming at the shortcomings of the traditional interference minimization algorithm,it is only a simple idea to calculate the maximum flow for all incoming and outgoing ports and determine the key link set.A resource allocation algorithm based on traffic prediction is proposed.According to the Markov prediction model of subtractive clustering,the changes of future traffic is predicted.Based on the maximum network flow value of the node pair,the weight of each link is calculated and used as the criterion of link resource importance The greater the weight,the more need to protect.At the same time,it is necessary to consider the delay of this attribute when the resource allocation based on traffic prediction is minimized.In this paper,we propose a resource allocation algorithm based on traffic prediction for interference and delay.The multi-objective optimization is transformed into single-objective optimization by fuzzy judgment,and then the path search is carried out by the improved heuristic search algorithm.