Research On Routing And Data Dissemination Algo-rithmsBased On Network Status And Behavior Pre-Diction In Cognitive Networks

Cognitive networks consist of a large number of intelligent nodes, which are equipped with sensing, data processing and communicating components. The cognitive nodes cooper-ate with each other to perceive the network environment, make behavioral decisions and re-source allocation according to the demands of the users and optimize performance and quality of service of the entire network. Due to the characteristics of intelligence, adaptiveness and self-management, cognitive networks have shown great potential in the application of many fields, such as social service, military communication, environment protection, intelligence transportation, disaster prediction and rescue and so on.Cognitive networks obtain the information of network status and behaviors, analyze and predict the future network status and behaviorsby learning and reasoning mechanisms, and make behavioral decisions according to the needs of the users and applications,whichcan op-timize the network performance and provide high-quality network services. At present, as human’s social life and network services are closely combined, social and community intelli-gence systems provide technologicalbasesfor social services.The system architecture and the design concept of cognitive networks provide the social intelligence systems with the effec-tive network model, and they can supplytechnical support and deployment schemes for the application systems of social and community intelligence.Based on the theory and architecture of cognitive networks and combined with the ap-plication system of social intelligence, this dissertation focuses on the adaptive routing schemes and data services of socialized applications in cognitive networks and perceives the information of the network status and parameters of users’behaviors. This dissertation ana-lyzes and predicts the environmental sensing information using the mathematical models of learning and reasoning mechanism, artificial intelligence algorithms, classical theories and methods of operational research. Based on the already noted, the adaptive routing algorithms, data dissemination mechanisms and related key technologies suitable for cognitive networks are studied in the dissertation.Firstly, the cognitive networking system is designed with the function of traffic predic-tion.The traffic prediction model is constructedaccording to the characteristic of cognitive networks. Depending on the traffic prediction model-MMSE, a routing algorithm, Mini-mum Workload Routing Algorithm (MWR) is proposedin cognitive networkswhich selects the route with the lightest traffic load,and the traffic load on each link of the routing pathis no more than the traffic threshold. Further more, we extend the scheme of MWR and propose the Adaptive Traffic Prediction Routing Algorithm (ATPRA) considering both traffic load and length load of the entire route and adaptively selecting a route of the lowest aggregated load by adjusting the threshold of traffic.Cognitive networks embody a sense of dynamic respon-siveness,since actions are typically taken in response to changing circumstances and changing resource availability to guaranty the end-to-end quality of service forusers.A multi-path rout-ing algorithm based on traffic prediction model, Efficient Traffic Aware Multi-path Routing (ETAMR) is proposed. ETAMR considers traffic distribution, node load and shortest path to build a multi-path routing, and selects the primary route with the shortest delay and lowest traffic load, meanwhile according to the real time traffic load it dynamically triggers the backup paths. ETAMR has good performance at network load balanceand can predict and avoid unexpected circumstances such as congestion and link failure of the network. The sim-ulation results illustrate thatour presented algorithms have good performances at load balanc-ing and lower transmission delay, which is validated by the simulation.Secondly, the cognitive network technology has been applied to the opportunistic net-works basedon social and community intelligence, and this dissertation aims at studying the mobile node location predication mechanism. The paper proposes a Social-relationship-based Mobile node Location Prediction algorithm (SMLP). The SMLP algorithm models applica-tion scenarios based on geographic locations, and extracts social relationships of mobile nodes from nodes’movement pattern. The SMLP algorithm preliminarily predicts the node’s mobility based on the Markov model, and then amends the prediction results using location information of other nodes which have strong relationship with the node. Two algorithms, SMLP1and SMLPN, are proposed based on the Markov model and the weighted Markov modelrespectively. Finally, the real data set are exploited for simulations. Simulation results show that SMLP acquires higher prediction accuracy than the Markov model, SMLPN achieves more accuracy on prediction than SMLP1, and obtains comparable prediction accu-racy with order-2Markov model while presents extra lower algorithm complexity.Thirdly, according to the opportunistic cognitive networks (OCNs), data dissemination services rely on the contact opportunity of mobile nodes. We propose three kinds ofdata dis-semination mechanisms applied in the OCNs. Ant Colony Optimizationbased DAta dissemi-nation (ACODAD) is proposed for fully distributed network environment. ACODAD uses the swarm intelligence mechanism, Ant Colony Optimization (ACO), to select the forwarding nodes according to the intimacy between the forwarding nodes and the destination node. The mechanism adopts cognitive heuristics technology based on ACO, and designsthe data dis-semination scheme based on self-adaptive data forwarding scheme in OCNs. LOcation Pre-diction based DAta Dissemination (LOPDAD) and LOcation-Prediction and Swarm-Intelligence based data dissemination (LOPSI) are suitable for a centralized and dis-tributed hybrid system. LOPDAD uses location prediction information to calculate the max-imum probability of the location where the forwarding nodes and the destination node en-counter and then completes data dissemination. The LOPSI algorithm is a probabilistic rout-ing protocol combining location prediction and the ant colony optimization. It firstly predicts possible locations of encountering nodes and destination(s) in successive time series. The mobile node compares the intimacybetween the destination node and potential relay nodes using ACO, and then makes a forwarding decision based on intimacy and the node mobili-ty.Meanwhile all the three algorithmspresented consider buffer management of the mobile node in order to to ensure operational efficiency and life cycle of the wholenetwork. Exten-sive simulationsare carried out on the real-world campus scenario to evaluate the proposed schemes in terms of transmission cost, delivery ratio, average hops and delivery latency, which achievebetter routing performances compared to the typical routing schemes in OCNs.Finally, the sustained participation and service reliability provided by the node are es-sential to the data collection and data dissemination services provided by OCNs for social and community intelligence systems. The participants consume their own resources to finish datatransmission services so it is necessary to provide correspondingincentive mecha-nismswhich can ensurethe enthusiasm and persistenceon participation of the nodes. To solve the issues mentioned above, this dissertation proposes a Reputation-based Participate Incen-tive Algorithm (RBPIA). RBPIA evaluates participants in terms ofdata reliability and bidding reliabilityto create a reputation model. The incentive mechanism based on such reputation model motivates participants to collect reliable data in opportunistic cognitive systems, while minimizing incentive cost for maintaining sufficient number of reliable participants. Simula-tions are conducted in different scenarios to test the performance of RBPIA. The results show that RBPIA remarkably increases the winning probability of participants who provide accu-rate data and reduces the cost for retaining sufficient number of participants.On the basis of RBPIA, we also designa data dissemination strategy to ensure the reliable data dissemination in OCNs.