Trust is a judgement of unquestionable utility — as humans we use it every day of our lives. However, trust has suffered from an imperfect understanding, a plethora of definitions, and informal use in the literature and in everyday life. It is common to say “I trust you, ” but what does that mean? This thesis provides a clarification of trust. We present a formalism for trust which provides us with a tool for precise discussion. The formalism is implementable: it can be embedded in an artificial agent, enabling the agent to make trust-based decisions. Its applicability in the domain of Distributed Artificial Intelligence (DAI) is raised. The thesis presents a testbed populated by simple trusting agents which substantiates the utility of the formalism. The formalism provides a step in the direction of a proper understanding and definition of human trust. A contribution of the thesis is its detailed exploration of the possibilities of future work in the area. Summary 1. Overview This thesis presents an overview of trust as a social phenomenon and discusses it formally. It argues that trust is: • A means for understanding and adapting to the complexity of the environment. • A means of providing added robustness to independent agents. • A useful judgement in the light of experience of the behaviour of others. • Applicable to inanimate others. The thesis argues these points from the point of view of artificial agents. Trust in an artificial agent is a means of providing an additional tool for the consideration of other agents and the environment in which it exists. Moreover, a formalisation of trust enables the embedding of the concept into an artificial agent. This has been done, and is documented in the thesis. 2. Exposition There are places in the thesis where it is necessary to give a broad outline before going deeper. In consequence it may seem that the subject is not receiving a thorough treatment, or that too much is being discussed at one time! (This is particularly apparent in the first and second chapters.) To present a thorough understanding of trust, we have proceeded breadth first in the introductory chapters. Chapter 3 expands, depth first, presenting critical views of established researchers.

We present a formal model of negotiation between autonomous agents. The purpose of the negotiation is to reach an agreement about the provision of a service by one agent for another. The model de nes a range of strategies and tactics that agents can employ to generate initial o ers, evaluate proposals and o er counter proposals. The model is based on computationally tractable assumptions, demonstrated in the domain of business process management and empirically evaluated. Keywords: Multi-agent systems, Negotiation, Business Process Management 1

When autonomous agents attempt to coordinate action, it is often necessary that they reach some kind of consensus. Reaching consensus has traditionally been dealt with in the Distributed Artificial Intelligence literature via negotiation. Another alternative is to have agents use a voting mechanism; each agent expresses its preferences, and a group choice mechanism is used to select the result. Some choice mechanisms are better than others, and ideally we would like one that cannot be manipulated by untruthful agents. Coordination of actions by a group of agents corresponds to a group planning process. We here introduce a new multi-agent planning technique, that makes use of a dynamic, iterative search procedure. Through a process of group constraint aggregation, agents incrementally construct a plan that brings the group to a state maximizing social welfare. At each step, agents vote about the next joint action in the group plan (i.e., what the next transition state will be in the eme...

In this paper we present three scheduling mechanisms that are manipulation-proof for closed systems. The amount of information that each user must encode in the mechanism increases with the complexity of the mechanism. On the other hand, the more complex the mechanism is, the more it maintains the privacy of the users. The first mechanism is a centralized, calendar-oriented one. It is the least computationally complex of the three, but does not maintain user privacy. The second is a distributed meeting-oriented mechanism that maintains user privacy, but at the cost of greater computational complexity. The third mechanism, while being the most complex, maintains user privacy (for the most part) and allows users to have the greatest influence on the resulting schedule. 1 Introduction The basic research problem in meeting scheduling is that of timing, that is, when to set a meeting. This question becomes more complicated when there are several meetings to be scheduled that involve the sa...

the rules by which agents in an encounter will interact. Once the rules of encounter have been determined, each builder of each agent is free to design his own machine any way that he wants. However, the rules that were established will certainly affect the choices he makes in building his own agent.

In this paper, we propose an algorithm which can improve Katz and Rosenschein's plan verification algorithm. First, we represent the plan-like relations with adjacency lists and inverse adjacency lists to replace adjacency matrixes. Then, we present a method to avoid generating useless sub-graphs while generating the compressed set. Last, we compare two plan verification algorithms. We not only prove that our algorithm is correct, but also prove that our algorithm is better than Katz and Rosenschein's algorithm both on time complexity and space complexity.

A negotiation mechanisms is presented for a real world problem of task distribution among a set of autonomous computational agents within a business process. The mechanism formally represents the protocol of agent interactions, the set of negotiation issues which agents negotiate over and the tactical and strategic reasoning procedures involved in negotiation decision making of an agent. Preliminary experimental evaluation of the model is also presented.

We present a formal model of negotiation between autonomous agents. The purpose of the negotiation is to reach an agreement about the provision of a service by one agent for another. The model defines a range of strategies and tactics that agents can employ to generate initial offers, evaluate proposals and offer counter proposals. The model is based on computationally tractable assumptions, demonstrated in the domain of business process management and empirically evaluated.

Abstract. Multiagent planning methods are concerned with planning by and for a group of agents. If the agents are selfinterested, they may be tempted to lie in order to obtain an outcome that is more rewarding for them. We therefore study the multiagent planning problem from a mechanism design perspective, showing how to incentivise agents to be truthful. We prove that the well-known truthful VCG mechanism is not always truthful in the context of optimal planning, and present a modi cation to x this. Finally, we present some (domain-dependent) poly-time planning algorithms using this x that maintain truthfulness in spite of their non-optimality. 1

When autonomous agents attempt to coordinate action, it is often necessary that they reach some kind of consensus. Reaching consensus has traditionally been dealt with in the Distributed Artificial Intelligence literature via negotiation. Another alternative is to have agents use a voting mechanism; each agent expresses its preferences, and a group choice mechanism is used to select the result. Some choice mechanisms are better than others, and ideally we would like one that cannot, be manipulated by untruthful agents. Coordination of actions by a group of agents corresponds to a group planning process. We here introduce a new multi-agent planning technique, that makes use of a dynamic, iterative search procedure. Through a process of group constraint aggregation, agents incrementally construct a plan that brings the group to a state maximizing social welfare. At each step, agents vote about the next joint action in the group plan (i.e., what the next transition state will be in the emerging plan) Using this technique agents need not fully reveal their preferences, and the set of alternative final states need not be generated in advance of a vote. With a minor variation, the entire procedure can be made resistant to untruthful agents. 1