Performance optimization based on non-additive measurement under heterogeneous network

By integrating different radio technologies, the heterogeneous wireless network is excepted to support end-to-end QoS and provide the "best" user experience. However these integrated networks have different access technologies and complex designed parameters. Current researches on the uncertain relationship among those parameters in such a dynamic and complicated system made it difficult to optimize the global network performance. As their inter-dependencies of each parameter may have a nonlinear relationship on the system performance metric of interest. This paper proposes a novel approach to analyze the global optimization problem in a heterogeneous network. We adopt a non-additive model based on Choquet integral to deal with the observed data in order to determine the network parameters' significance. By adjusting the most significant parameter, we can optimize network transfer performance. A simulation platform, which is deployed two applications such as voice and video conferencing, has been developed using OPNET to support our approach. Experimental evaluation shows that the proposed method's feasibility and effectiveness.