"Negotiation StationInteresting. And I tried unsuccessfully to access the "landmark 'Gaia' paper on agent-oriented analysis and design". I hate getting a teaser like this, and not being able to review the article itself, without having to pay for a subscription. But I was able to find this expanation in this paper, Semantic treatment of queries based on agents to improve information retrieval in federated digital libraries by Student: María Auxilio Medina Nieto with Ph.D. Alfredo Sánchez acting as Advisor:
Jennings -- a computer science professor at the University of Southampton -- assesses the effectiveness of so-called 'argumentation-based negotiation' (ABN) for computer agents in a recently published paper.
Agents are computer systems to which an operator can delegate tasks. Considered autonomous in comparison to programs that depend on every keystroke, agents are increasingly used in a wide range of industrial and commercial domains, including robotics, e-commerce, computer games Latest News about computer games and information retrieval.
In systems with more than one agent, where 'autonomous entities pursue their own goals, conflict is inevitable,' Jennings explained.
Negotiation among the agents is the best way to 'resolve these problems,' he added.
To resolve conflicts through negotiation, computers need artificial intelligence Latest News about artificial intelligence programs, which are 'increasingly being used on the Internet, in our homes, and in the workplace,' Jennings told NewsFactor.
'To improve their performance, we need to ensure they have the ability to overcome real-world problems, such as conflict,' he stressed.
'I am very much in agreement with Prof. Jennings on the importance and the promise of agent technology,' said Agentis chief technology officer David Kinny, who with Jennings and Michael Wooldridge co-authored the landmark 'Gaia' paper on agent-oriented analysis and design.
'Negotiation techniques are crucial in open-agent systems,' said Kinny, 'where agents representing different individuals or organizations interact -- as well as in any systems where agents have conflicting goals or information.'"
4.1.2 Gaia methodologyOkay, still above my head, let me know if you find an explanation that's a little easier to come to grips with, and maybe why the 'Gaia' paper on agent-oriented analysis and design was considered 'landmark'.
Gaia is a top−down methodology for agent−oriented analysis and design. It is based on the
concept of roles. A multi−agent system is understood as a computational organization
formed by the interaction of various roles. To understand and model a complex system, a
set of two agent−specific types of concepts are provided: abstract concepts, (roles,
permissions, responsibilities, protocols, activities, aliveness properties and safety
properties), and concrete concepts, (agent types, services, acquaintances). These concepts
are briefly described in this section.
The conceptual structure of Gaia has three stages: requirements statement, analysis
and design. Requirements statement can be done with traditional software engineering
techniques. Models form the other two stages. The analysis stage is composed by roles
model and interactions model, whereas the design stage includes agent model, services
model and acquaintance model. Requirements statement operates as input to the analysis
stage. Each model of this stage is related with the models of the design stage [Wooldridge
et al. 2000].
The understanding of the system and its components without implementation details
is done at the analysis stage. The most important concepts at this level are roles and
interactions; each one constitutes a model. In the roles model, a role is defined by the
following attributes:• Responsibilities to determine the functionality. They are subdivided in aliveness
responsibilities or safety responsibilities according to if they describe states of
affairs that an agent must bring about or acceptable states of affairs maintained
across the whole execution, respectively.
• Permissions or rights associated with a role.
• Activities that are performed by an agent without interacting with other agents.
• Protocols to define how a role interacts with other roles.
These components form a template called role schemata. Roles model consists of a
set of role schemata.