The ‘Always Best Connected’ vision is built around the scenario of a mobile user seamlessly roaming within a multi-operator multi-technology multi-terminal multi-applicationudmulti-user environment supported by the next generation of wireless networks. In this heterogeneous environment, users equipped with multi-mode wireless mobile devices willudaccess rich media services via one or more access networks. All these access networks may differ in terms of technology, coverage range, available bandwidth, operator, monetary cost, energy usage etc. In this context, there is a need for a smart network selection decision to be made, to choose the best available network option to cater for the user’s current application and requirements. The decision is a difficult one, especially given the number and dynamics of the possible input parameters. What parameters are used and how those parameters model the application requirements and user needs is important. Also, game theory approaches can be used to model and analyze the cooperative or competitive interaction between the rational decision makers involved, which are users, seeking to get good service quality at good value prices, and/or the network operators, trying to increase their revenue.udThis thesis presents the roadmap towards an ‘Always Best Connected’ environment. The proposed solution includes an Adapt-or-Handover solution which makes use of a SignaludStrength-based Adaptive Multimedia Delivery mechanism (SAMMy) and a Power-Friendly Access Network Selection Strategy (PoFANS) in order to help the user in takinguddecisions, and to improve the energy efficiency at the end-user mobile device. A Reputation-based System is proposed, which models the user-network interaction as a repeated cooperative game following the repeated Prisoner’s Dilemma game from Game Theory. It combines reputation-based systems, game theory and a network selection mechanism in order to create a reputation-based heterogeneous environment. In this environment, the users keep track of their individual history with the visited networks. Every time, a user connects to a network the user-network interaction game is played. The outcome of the game is a network reputation factor which reflects the network’s previous behavior in assuring service guarantees to the user. The network reputation factor will impact the decision taken by the user next time, when he/she will have to decide whether to connect or not to that specific network. The performance of the proposed solutions was evaluated through in-depth analysis and both simulation-based and experimental-oriented testing. The results clearly show improved performance of the proposed solutions in comparison with other similar state-of-the-art solutions. An energy consumption study for a Google Nexus One streaming adaptive multimedia was performed, and a comprehensive survey on related Game Theory research are provided as part of the work.
展开▼
