Research On The Network Selection Algorithm In Heterogeneous Wireless Networks

The next-generation wireless network will be the heterogeneous wireless networks (HWNs) which includes varieties of wireless networks, conventional or emerging in the future. It will aim to provide various services for subscribers with the help of the feature of the mobile nodes (MN) which have multiple interfaces and the heterogeneity between the varieties of wireless networks. However, ratio resource management (RRM) in the HWNs has always faced a series of problems and challenges, which mobile terminals have to deal with first is that the network selection in heterogeneous wireless networks, and research on network selection algorithm has become a very important issue. The thesis deeply discusses and studies the network selection algorithms from the aspects of multi-attribute decision algorithm and its performance, performance analysis based on MDP, OWA operator which applied in network selection scenarios.The thesis studies and simulates seven kinds of multi-attribute decision making algorithms, such as Simple Additive Weight (SAW), Multiplicative Exponent Weight (MEW), Technique for Order Preference by Similarity to Ideal Solution Algorithm (TOPSIS), Grey Relational Analysis (GRA), Weighted Markov Chain (WMC), Elimination and Choice Translating Priority (ELECTRE), and analyzes the performance of the algorithms. In terms of the results of network selection, SAW, MEW, TOPSIS and VIKOR algorithm provides a more conservative classification, however, ELECTRE, GRA and WMC algorithm can be more comprehensive for the situation which has more additional factors.For a more comprehensive analysis of the performance of heterogeneous network selection algorithm, this thesis proposed a 4-Dimension Markov model considering both voice and data users in heterogeneous networks consisting of GSM/EDGE (RAT based on TDMA) and UMTS (RAT based on WCDMA). And variables named state weight are added in the state transfer process whose values corresponding to different network selection algorithms. Then the paper also showed the performance analysis and comparison of Random Selection Policy, Service-Based 1 Selection Policy, Service-Based 2 Selection Policy, Load Balance Selection Policy by using the 4D Markov model. Model parameters are determined according to system model and handoff rate analysis for heterogeneous networks. And successive over-relaxation iteration method is used to solve the steady-state distribution for Markov model. Finally, the closed formulas of performance parameters are obtained for the proposed algorithm, e.g. call blocking probability, handoff blocking probability, throughput. The simulation results show that the 4D Markov model used in performance analysis of network selection algorithms is reasonable.OWA operator has been widely used in the field of fuzzy logic, multi-attribute decision making and group decision-making. This thesis applies OWA operator to the scene of network selection and proposes a heterogeneous network selection algorithm based on OWA operator in heterogeneous networks consisting of Wireless Local Area Networks (WLAN), Worldwide Interoperability for Microwave Access (WiMAX) and Universal Mobile Telecommunications System (UMTS). This algorithm first normalizes the network attributes of each candidate network, and then selects one of the "biggest factor" to compute its weight with the non-neat aggregation method, finally computes the OWA values of each candidate network by using the defined formula of OWA operator and chooses the best network. By comparing with the SAW, TOPSIS, GRA, the simulation results show that the heterogeneous network selection algorithm based on OWA operator is reasonable.