本論文著眼於All-IP網路中的預算分配。我們定義網路的服務品質只與各使用者使用網路頻寬的要求有關，提出一種使管理者能以統計上百分比來估計網路服務品質的方法。這個方法包含路徑選擇以及頻寬分配二個階段。為了展現這種方法的可行性，我們列舉一些數據來分別比較以最大滿意度和最小成本為目標的不同分配結果，作為使用這個方法的參考。 / In this thesis, we focus on budget allocation for All-IP networks. We propose a method which assists managers to estimate the quality of service on networks. The quality of service on networks is defined by satisfaction functions that are simply written in terms of bandwidth required by the users on the network. We present a two-phase approach which includes a path se-lection and a scheme for bandwidth allocation. In order to illustrate an easy implementation of this approach, we also develop the Maximum Satisfaction Method and the Minimum Cost Method. Numerical examples are given to show the effectiveness of our approach.

This process developed in this work, Strategy Optimization for the Allocation of Resources (SOAR), is a strategic planning methodology based off Integrated Product and Process Development and systems engineering techniques. Utilizing a top down approach, the process starts with the creation of the organization vision and its measures of effectiveness. These measures are prioritized based on their application to external world scenarios which will frame the future. The programs which will be used to accomplish this vision are identified by decomposing the problem. Information is gathered on the programs as to the application, cost, schedule, risk, and other pertinent information. The relationships between the levels of the hierarchy are mapped utilizing subject matter experts. These connections are then utilized to determine the overall benefit of the programs to the vision of the organization. Through a Multi-Objective Genetic Algorithm a tradespace of potential program portfolios can be created amongst which the decision maker can allocate resources. The information and portfolios are presented to the decision maker through the use of a Decision Support System which collects and visualizes all the data in a single location. This methodology was tested utilizing a science and technology planning exercise conducted by the Unites States Navy. A thorough decomposition was defined and technology programs identified which had the potential to provide benefit to the vision. The prioritization of the top level capabilities was performed through the use of a rank ordering scheme and a previous naval application was used to demonstrate a cumulative voting scheme. Voting was performed utilizing the Nominal Group Technique to capture the relationships between the levels of the hierarchy. Interrelationships between the technologies were identified and a MOGA was utilized to optimize portfolios with respect to these constraints and information was placed in a DSS. This formulation allowed the decision makers to assess which portfolio could provide the greatest benefit to the Navy while still fitting within the funding profile.

In this thesis, we examine stochastic techniques for overcoming game theoretic and computational issues in the collective decision making process of self-interested individuals. In particular, we examine truthful, stochastic mechanisms, for settings with a strong budget balance constraint (i.e. there is no net flow of money into or away from the agents). Building on past results in AI and computational social choice, we characterise affine-maximising social choice functions that are implementable in truthful mechanisms for the setting of heterogeneous item allocation with unit demand agents. We further provide a characterisation of affine maximisers with the strong budget balance constraint. These mechanisms reveal impossibility results and poor worst-case performance that motivates us to examine stochastic solutions. To adequately compare stochastic mechanisms, we introduce and discuss measures that capture the behaviour of stochastic mechanisms, based on techniques used in stochastic algorithm design. When applied to deterministic mechanisms, these measures correspond directly to existing deterministic measures. While these approaches have more general applicability, in this work we assess mechanisms based on overall agent utility (efficiency and social surplus ratio) as well as fairness (envy and envy-freeness). We observe that mechanisms can (and typically must) achieve truthfulness and strong budget balance using one of two techniques: labelling a subset of agents as ``auctioneers'' who cannot affect the outcome, but collect any surplus; and partitioning agents into disjoint groups, such that each partition solves a subproblem of the overall decision making process. Worst-case analysis of random-auctioneer and random-partition stochastic mechanisms show large improvements over deterministic mechanisms for heterogeneous item allocation. In addition to this allocation problem, we apply our techniques to envy-freeness in the room assignment-rent division problem, for which no truthful deterministic mechanism is possible. We show how stochastic mechanisms give an improved probability of envy-freeness and low expected level of envy for a truthful mechanism. The random-auctioneer technique also improves the worst-case performance of the public good (or public project) problem. Communication and computational complexity are two other important concerns of computational social choice. Both the random-auctioneer and random-partition approaches offer a flexible trade-off between low complexity of the mechanism, and high overall outcome quality measured, for example, by total agent utility. They enable truthful and feasible solutions to be incrementally improved on as the mechanism receives more information and is allowed more processing time. The majority of our results are based on optimising worst-case performance, since this provides guarantees on how a mechanism will perform, regardless of the agents that use it. To complement these results, we perform empirical, average-case analyses on our mechanisms. Finally, while strong budget balance is a fixed constraint in our particular social choice problems, we show empirically that this can improve the overall utility of agents compared to a utility-maximising assignment that requires a budget imbalanced mechanism.

THE AUSTRALIAN HEALTH CARE SYSTEM has an overarching objective to improve the well-being of all Australians in an equitable and efficient manner. But like most developed economy health care systems, it has experienced a continual increase in demand for health care services along with increased pressure to improve efficiency, quality, and sustainability. To assist in health sector management, policy formulation, investment decisions and reform, the Australian government developed the National Health Performance Framework (NHPF). The NHPF employs performance indicators across nine dimensions of health care, including Effectiveness, Appropriateness, Efficiency, Responsiveness, Accessibility, Safety, Continuity, Capability, and Sustainability. While the National Health Performance Committee has recognised that performance indicators used within the NHPF are inadequate, this thesis argues that the solution is not a simple matter of collecting additional data and constructing new and ???improved??? indicators. Due to resource constraints within the health care system there is an implicit performance trade-off across dimensions. The NHPF must take into consideration the value individuals place on the health care dimensions to enable a shift of limited resources to those areas that are most valued. The starting point for the NHPF should be to determine what society wants out of a health system. The purpose of this thesis is to determine Australian society???s preferences for performance across the nine NHPF dimensions of health care. This is achieved using a choice modelling experiment, which describes the performance of the current health care system and alternative health care systems the government could work towards, and asks respondents to compare and choose which system they prefer. A mixed multinomial logit model is used to analyse respondent choices in order to incorporate alternative tastes across attributes, and correlation of tastes across alternatives and scenarios. Relative values attached to the nine NHPF dimensions of health care are calculated and preferences for the dimensions are ranked. The thesis concludes by exploring individual preferences derived form the choice modelling experiment in the context of social welfare theory. It also outlines the strengths and weaknesses of the methodology, provides suggestions for further research, and offers a use for social preferences in the development of performance frameworks within the Australian health care system.

The focus of this thesis is on the resource allocation problems in wireless sensor and cooperative networks. Typically, wireless sensor networks operate with limited energy and bandwidth are often required to meet some specified Quality-of-Service (QoS) constraints. The ultimate objective for the majority of the problems considered in this thesis is to save battery energy and maximize the network lifetime. / In the first part of this thesis, we employ complex mathematical models to emulate a variety of power drains in wireless sensor nodes. In the first instance, we address a lifetime optimization problem of a wireless TDMA/CDMA sensor network for joint transmit power and rate allocations. The effect of fast fading is captured by including rate outage and link outage constraints on each link. After that, a single-hop wireless sensor network is deployed for a certain application - to estimate a Gaussian source within a pre-specified distortion threshold. In this part, we consider lifetime maximization, in different multiple access protocols such as TDMA, an interference limited non-orthogonal multiple access (NOMA) and an idealized Gaussian multiple access channel. This problem is further studied in a multi-hop scenario where sensing and receiving powers are also included in addition to transmission power. Finally, we investigate a balancing problem between the source coding and transmission power for video wireless sensor systems where the sensor node is required to send the collected video clips, through wireless media, to a base station within a corresponding distortion threshold. All these energy saving and lifetime optimization problems in sensor networks can be formulated via nonlinear nonconvex optimization problems, which are generally hard to solve. However, with favourable variable substitution and reasonable approximation, most of these problems are shown to be convex. The only exception is the Gaussian source esitmation problem in NOMA scenario for which we provide a simple successive convex approximation based algorithm for the NOMA case that converges fast to a suboptimal solution. / In the second part of the thesis, we propose an optimal power allocation scheme with a K-block coding delay constraint on data transmission using a three node cooperative relay network assuming a block fading channel model. Channel information is fed back to the transmitter only in a causal fashion, so that the optimal power allocation strategy is only based on the current and past channel gains. We consider the two simplest schemes for information transmission using a three node (a source, a relay and a destination) relay network, namely the amplify and forward (AF) and decode and forward (DF) protocols. We use the dynamic programming methodology to solve the (K-block delay constrained) expected capacity maximization problem and the outage probability minimization problem with a short term sum power (total transmission power of the source and the relay) constraint. / The main contribution of the thesis is a comprehensive suite of power minimization and lifetime maximization methods that can be used in wireless sensor networks. We present several such applications and extensive numerical examples at the end of each chapter.

Dissertation (DPhil)--University of Stellenbosch, 2006. / Bibliography. Also available via the Internet.

Diss. (sammanfattning) Skara : Sveriges lantbruksuniv., 2002. / Härtill 4 uppsatser.

Diss. Stockholm : Karol. inst., 2003.

Thesis (Ph. D.)--University of Hong Kong, 2008. / Includes bibliographical references (leaf 110-117) Also available in print.

Thesis (Ed.D)--University of Delaware, 2007. / Principal faculty advisor: Douglas A. Archbald, School of Education. Includes bibliographical references.