Search

In the past decades, mobile traffic generated by devices such as smartphones, iphones, laptops and mobile gateways has been growing rapidly. While traditional direct connection techniques evolve to provide better access to the Internet, a new type of wireless network, mobile ad hoc network (MANET), has emerged. A MANET differs from a direct connection network in the way that it is multi-hopping and self-organizing and thus able to operate without the help of prefixed infrastructures. However, challenges such dynamic topology, unreliable wireless links and resource constraints impede the wide applications of MANETs. Routing in a MANET is complex because it has to react efficiently to unfavourable conditions and support traditional IP services. In addition, Quality of Service (QoS) provision is required to support the rapid growth of video in mobile traffic. As a consequence, tremendous efforts have been devoted to the design of QoS routing in MANETs, leading to the emergence of a number of QoS support techniques. However, the application independent nature of QoS routing protocols results in the absence of a one-for-all solution for MANETs. Meanwhile, the relative importance of QoS metrics in real applications is not considered in many studies. A Best Effort QoS support (BEQoS) routing model which evaluates and ranks alternative routing protocols by considering the relative importance of multiple QoS metrics is proposed in this thesis. BEQoS has two algorithms, SAW-AHP and FPP for different scenarios. The former is suitable for cases where uncertainty factors such as standard deviation can be neglected while the latter considers uncertainty of the problems. SAW-AHP is a combination of Simple Additive Weighting and Analytic Hierarchical Process in which the decision maker or network operator is firstly required to assign his/her preference of metrics with a specific number according to given rules. The comparison matrices are composed accordingly, based on which the synthetic weights for alternatives are gained. The one with the highest weight is the optimal protocol among all alternatives. The reliability and efficiency of SAW-AHP are validated through simulations. An integrated architecture, using evaluation results of SAW-AHP is proposed which incorporates the ad hoc technology into the existing WLAN and therefore provides a solution for the last mile access problems. The protocol selection induced cost and gains are also discussed. The thesis concludes by describing the potential application area of the proposed method. Fuzzy SAW-AHP is extended to accommodate the vagueness of the decision maker and complexity of problems such as standard deviation in simulations. The fuzzy triangular numbers are used to substitute the crisp numbers in comparison matrices in traditional AHP. Fuzzy Preference Programming (FPP) is employed to obtain the crisp synthetic weight for alternatives based on which they are ranked. The reliability and efficiency of SAW-FPP are demonstrated by simulations.

Mixed ionic and electronic conductor oxides, in particular La1-xSrxCoyFe1-yO3-d (LSCF), have been widely used as the cathode materials in solid oxide fuel cells for high-temperature energy applications. The focus of this thesis is primarily on constructing the instability phase diagram of Sr segregations on LSCF surfaces at the experimentally relevant temperatures and oxygen partial pressures using the first-principles density functional theory (DFT). A generic first-principles free-energy functional is developed to obtain the nonstoichiometric oxygen vacancy concentrations for the bulk and surface phases. These results agree well with the corresponding thermo-gravimetry measurements, and furthermore suggest that the oxygen vacancies are energetically stabilized at surfaces for all temperatures and oxygen partial pressures, while such surface stabilization effects become stronger at higher temperatures and lower oxygen partial pressures. Based on these nonstoichiometric oxygen vacancy predictions, we construct the free-energy phase diagrams of the Sr-segregation reaction as a function of temperature, oxygen partial pressure, and CO2 partial pressure for both the bulk and surface LSCF phases. Our results suggest that Sr segregations strongly accumulate towards the LSCF surface phase where the oxygen vacancy nonstoichiometries are abundant. Our results also indicate that the Sr segregation reactions are significantly enhanced at high temperatures, low oxygen partial pressures, and high CO2 partial pressures. The computed reaction temperature ranges are consistent with the total reflection X-ray fluorescence (TXRF) measurements.