Multi-Attribute Decision Algorithm Of Vertical Handoff In Heterogeneous Wireless Interworked Networks

The primary characteristic of next-generation communication system is that the various access technologies exist simultaneously, complement with each other and integrate seamlessly. Neither of the network meets high bandwidth, low latency, wide coverage and other requirements. Consequently, it is an inevitable trend that wireless networks will integrate with each other. Vertical handoff is the key technology of the integration of heterogeneous wireless networks. Efficient and accurate vertical handoff algorithm is the urgent requirement of the integration of heterogeneous wireless networks.Vertical handoff algorithms based on quality of service don't establish global QoS factor evaluation system. Some of the Qos factors exist interdependent and overlapping information, which will lead to an incorrect handoff decision. In order to improve handoff decision accuracy, some algorithms based on artificial intelligence was proposed, such as grey prediction and fuzzy logic. But the computation complexity of the artificial intelligence is very large. Besides, it doesn't solve the problem of the overlapping information exist between QoS factors.In order to solve the problem of the traditional methods, we propose a new vertical handoff algorithm based on principle component analysis (PCA), which optimize the traditional methods through establishing QoS factor evaluation system and eliminating the overlapping information.PCA transforms vertical handoff algorithm into multi-attribute decision making problem and constructs multi-attribute decision making model. Besides, PCA establish QoS factor evaluation system from network availability, timeliness, security, reliability and so on. AHP is responsible for deriving the weights of QoS parameters. Based on this, we propose an efficient algorithm to extract the QoS principal component and make handoff decision.The performance of comparison between the PCA and existing scheme shows that the proposed scheme have high efficiency, accurate decision, low computational complexity. First, PCA extracts the QoS principle component based on the QoS principle component contribution rate, which improve the efficiency of the algorithm. Besides, PCA eliminates the overlapping information between QoS factors. Furthermore, the QoS principal component extracted from PCA could be considered as the input of other handoff algorithm, such as handoff predictive algorithm, GRA and so on, which will efficiently reduce the computational complexity.Finally, based on the proposed method, we design and implement the handoff system of 3G-WLAN integrated network. The experiment result shows that PCA could help terminal to choose the best access network.