Current research in multi-agent systems (MAS) has advanced to the development of open MAS, which are characterized by the heterogeneity of agents, free exit/entry and decentralized control. Conflicts of interest among agents are inevitable, and hence automated negotiation to resolve them is one of the promising solutions. This thesis studies three modifications on alternating-offer bargaining protocols for automated negotiation in open MAS. The long-term goal of this research is to design negotiation protocols which can be easily used by intelligent agents in accommodating their need in resolving their conflicts. In particular, we propose three modifications: allowing non-monotonic offers during the bargaining (non-monotonic-offers bargaining protocol), allowing strategic delay (delay-based bargaining protocol), and allowing strategic ignorance to augment argumentation when the bargaining comprises argumentation (ignorance-based argumentation-based negotiation protocol). <p>Utility theory and decision-theoretic approaches are used in the theoretical analysis part, with an aim to prove the benefit of these three modifications in negotiation among myopic agents under uncertainty. Empirical studies by means of computer simulation are conducted in analyzing the cost and benefit of these modifications. Social agents, who use common human bargaining strategies, are the subjects of the simulation. <p>In general, we assume that agents are bounded rational with various degrees of belief and trust toward their opponents. In particular in the study of the non-monotonic-offers bargaining protocol, we assume that our agents have diminishing surplus. We further assume that our agents have increasing surplus in the study of delay-based bargaining protocol. And in the study of ignorance-based argumentation-based negotiation protocol, we assume that agents may have different knowledge and use different ontologies and reasoning engines. <p>Through theoretical analysis under various settings, we show the benefit of allowing these modifications in terms of agents expected surplus. And through simulation, we show the benefit of allowing these modifications in terms of social welfare (total surplus). Several implementation issues are then discussed, and their potential solutions in terms of some additional policies are proposed. Finally, we also suggest some future work which can potentially improve the reliability of these modifications.

Unsolicited e-mail has brought much annoyance to users, thus, making e-mail less reliable as a communication tool. This has happened because current email architecture has key limitations. For instance, while it allows senders to send as many messages as they want, it does not provide adequate capability to recipients to prevent unrestricted access to their mailbox. This research develops a new approach to equip recipients with ability to control access to their mailbox.<p>This thesis builds an ownership-based approach to control mailbox usage employing the CyberOrgs model. CyberOrgs is a model that provides facilities to control resources in multi-agent systems. We consider a mailbox to be a precious resource of its owner. Any access to the resource requires its owner's permission. Thus, we give recipients a capability to manage their valuable resource - mailbox. In our approach, message senders obtain a permission to send messages through negotiation. In this negotiation, a sender makes a proposal and the intended recipient evaluates the proposal according to their own policies. A sender's desired outcome of a negotiation is a contract, which conducts the subsequent communication between the sender and the recipient. Contracts help senders and recipients construct a long-term relationship.<p>Besides allowing individuals to control their mailbox, we consider groups, which represent organizations in human society, in order to allow organizations to manage their resources including mailboxes, message sending allowances, and contracts.<p>A prototype based on our approach is implemented. In the prototype, policies are separated from the mechanisms. Examples of policies are presented and a public policy interface is exposed to allow programmers to develop custom policies. Experimental results demonstrate that the system performance is policy-dependent. In other words, as long as policies are carefully designed, communication involving negotiation has minimal overhead compared to communication in which senders deliver messages to recipients directly.

Like many North American agricultural regions, Saskatchewan has experienced significant fundamental structural changes in farming. Structural change encompasses evolution in distribution of farm sizes, land tenure and financial characteristics, as well as variations in demographic and production characteristics. These issues are often a source of discontent among farm populations as it implies these populations are forced to adapt in a number of potentially unpleasant ways. These changes have profound and sometimes poorly understood effects on the rural economy for example, structural change affects rural population and therefore demand for rural infrastructure. <p>Traditional agricultural farm level analysis is often conducted using a representative farm or group, but this framework cannot capture the growing heterogeneity of modern farm operators or the current operating environment in agricultural regions. Farm profiles vary by demographic characteristics, such as age and education, and resource endowments. Agent-based simulation captures this heterogeneity through a farm by farm analysis, where after initialization, the regional economy evolves over time.<p>A synthetic population is created based on survey data and the land characteristics based on the actual land data in CAR 7B of Saskatchewan. A number of different price and yield time paths were created using a bootstrap procedure on historical data and the model evolved to potential agriculture structures that may occur in the model region, 30 years in the future.<p>Structural change occurs endogenously as farms interact in land markets, and make decisions on land use. Agents compete for available land in a purchase and lease market with land selling to the highest bidder. The dynamic nature of agent-based models allows individual farms to adjust land use in response to changing economic conditions and individual preferences. How individuals organize their resources will be critical to farm survival and growth.<p>The results indicate that many of the trends are the same under the different price and yield time paths, however the rate of change is significantly impacted by the price and yield time path that occurs. The model predicted the trend to fewer and larger farms will continue into the future. The forecasted distribution of smaller farms will decline and proportion of large farms will increase, while mid sized farms will remain relatively unchanged. The proportion of mixed farms, land use, and total livestock numbers depend significantly on the price and yield time path. The actual structure that will occur will be the result of the actual individual price and yield time path that occurs.

As popularity of online video grows, a number of models of advertising are emerging. It is typically the brokers usually the operators of websites who maintain the balance between content and advertising. Existing approaches focus primarily on personalizing advertisements for viewer segments, with minimal decision-making capacity for individual viewers. We take a resource ownership view on this problem. We view consumers attention space, which can be abstracted as a display screen for an engaged viewer, as precious resource owned by the viewer. Viewers pay for the content they wish to view in dollars, as well as in terms of their attention. Specifically, advertisers may make partial payment for a viewers content, in return for receiving the viewers attention to their advertising. Our approach, named FlexAdSense, is based on CyberOrgs model, which encapsulates distributed owned resources for multi-agent computations.<p> We build a market of viewers attention space in which advertisers can trade, just as viewers can trade in a market of content. We have developed key mechanisms to give viewers flexible control over the display of advertisements in real time. Specific policies needed for automated negotiations can be plugged-in. This approach relaxes the exclusivity of the relationship between advertisers and brokers, and empowers viewers, enhancing their viewing experience.<p> This thesis presents the rationale, design, implementation, and evaluation of FlexAdSense. Feature comparison with existing advertising mechanisms shows how FlexAdSense enables viewers to control with fine-grained flexibility. Experimental results demonstrate the scalability of the approach, as the number of viewers increases. A preliminary analysis of user overhead illustrates minimal attention overhead for viewers as they customize their policies.

Magnetic resonance imaging (MRI) is one of the most powerful imaging techniques for research and clinical diagnosis. To expand upon the intrinsic capabilities of MRI, new contrast agents that can detect the presence of biomarkers in vivo are being developed. My Masters thesis research focuses on the design and synthesis of a new MRI contrast agent that can detect glutathione (GSH), a biomarker that has been implicated in a number of oxidative stress diseases. This new MRI contrast agent is based on chelated dysprosium (Dy), an inorganic metal, which provides negative contrast to surrounding tissue. Preliminary data has shown that attaching a poly(ethylene glycol) (PEG) chain to the Dy chelate, effectively increasing its molecular weight, enhances the contrast ability of Dy. Using this new information, the contrast agent was designed to have a large molecular weight PEG chain attached to the Dy chelate through a disulfide, creating a thiol-sensitive linkage. In the presence of a thiol-containing molecule such as GSH, the Dy will be dePEGylated through a disulfide exchange reaction, removing the molecular weight effect of the PEG, and allowing for the detection of GSH by MRI. This new MRI contrast agent could provide insight into the progression and diagnosis of oxidative stress pathologies associated with GSH.

My study examines the relationship between principal and agent in a moral hazard setting where the principal has the ability to monitor the actions of the agent at an interim stage of the project. I show that monitoring can induce the agent to exert higher levels of effort and can result in a reallocation of project payoffs between the two parties. This reallocation is not a one-way street: Situations exist where monitoring encourages greater effort from the agent, resulting in greater project payoffs for both principal and agent. For projects that are characterized as high-risk, high-reward projects where agent involvement is costly, monitoring is often the optimal strategy; this is an explanation for why venture capital type investments are the subject of intense monitoring. When the principal can share monitoring results at an interim stage with the agent, the agent is able to modify his effort levels in certain situations for the benefit of both parties.

Reinforcement learning is a machine learning technique designed to mimic the way animals learn by receiving rewards and punishment. It is designed to train intelligent agents when very little is known about the agent’s environment, and consequently the agent’s designer is unable to hand-craft an appropriate policy. Using reinforcement learning, the agent’s designer can merely give reward to the agent when it does something right, and the algorithm will craft an appropriate policy automatically. In many situations it is desirable to use this technique to train systems of agents (for example, to train robots to play RoboCup soccer in a coordinated fashion). Unfortunately, several significant computational issues occur when using this technique to train systems of agents. This dissertation introduces a suite of techniques that overcome many of these difficulties in various common situations. First, we show how multi-agent reinforcement learning can be made more tractable by forming coalitions out of the agents, and training each coalition separately. Coalitions are formed by using information-theoretic techniques, and we find that by using a coalition-based approach, the computational complexity of reinforcement-learning can be made linear in the total system agent count. Next we look at ways to integrate domain knowledge into the reinforcement learning process, and how this can signifi-cantly improve the policy quality in multi-agent situations. Specifically, we find that integrating domain knowledge into a reinforcement learning process can overcome training data deficiencies and allow the learner to converge to acceptable solutions when lack of training data would have prevented such convergence without domain knowledge. We then show how to train policies over continuous action spaces, which can reduce problem complexity for domains that require continuous action spaces (analog controllers) by eliminating the need to finely discretize the action space. Finally, we look at ways to perform reinforcement learning on modern GPUs and show how by doing this we can tackle significantly larger problems. We find that by offloading some of the RL computation to the GPU, we can achieve almost a 4.5 speedup factor in the total training process.

none

none

Bargaining is a frequent activity in the shopping process, and it becomes a trend in electronic trading. In order to facilitate the on-line automatic bargaining activity, we develop three algorithms on the multi-agent system in this thesis. The first algorithm is the pattern generalization algorithm used for generalizing common patterns from transaction records. The second one is the pattern matching algorithm used on-line for identifying possible bargaining patterns from the pattern bases. To deal with the situation that there is no matched pattern, we design the dynamic price issuing algorithm using the utility theory to determine the seller¡¦s price and the timing a deal should be closed. We conducted a series of field experiments to evaluate the proposed algorithms on different seller¡¦s risk perspectives and compared the performance with conventional bargaining methods. The results show that the proposed methods obtain encouraging performance. The major contribution of this research is the initiation efforts on developing data mining algorithms for facilitating the price bargaining process for e-commerce.