Designing Agent-Based Process Systems-Extending the OPEN Process Framework

Originally a development methodology targeted at object technology, the OPEN Process Framework (OPF) is found to be a successful basis for extensions that support agent-oriented software development. Here we describe the process components necessary to agent-oriented support and illustrate the extensions by means of two small case studies that illustrate the extensions by means of two small case studies that illustrate both task-driven processes and goal-driven processes. The additional process components for Tasks and Techniques are all generated from the OPF’s metamodel, which gives the OPF its flexibility and tailorability to a wide variety of situations—here agent-orientation.

Evolutionary Computation as a Paradigm for Engineering Emergent Behaviour in Multi-Agent Systems

In the multi-agent system (MAS) context, the theories and practices of evolutionary computation (EC) have new implications, particularly with regard to engineering and shaping system behaviors. Thus, it is important that we consider the embodiment of EC in “real” agents, that is, agents that involve the real restrictions of time and space within MASs. In this chapter, we address these issues in three ways. First, we relate the foundations of EC theory to MAS and consider how general interactions among agents fit within this theory. Second, we introduce a platform independent agent system to assure that our EC methods work within the generic, but realistic, constraints of agents. Finally, we introduce an agent-based system of EC objects. Concluding sections discuss implications and future directions.

Engineering Emotionally Intelligent Agents

Interacting with intelligence in an ever-changing environment calls for exceptional performances from artificial beings. One mechanism explored to produce intuitive-like behavior in artificial intelligence applications is emotion. This chapter examines the engineering of a mechanism that synthesizes and processes an artificial agent’s internal emotional states: the Affective Space. Through use of the affective space, an agent can predict the effect certain behaviors will have on its emotional state and, in turn, decide how to behave. Furthermore, an agent can use the emotions produced from its behavior to update its beliefs about particular entities and events. This chapter explores the psychological theory used to structure the affective space, the way in which the strength of emotional states can be diminished over time, how emotions influence an agent’s perception, and the way in which an agent can migrate from one emotional state to another.

Toward an Organization-Oriented Design Methodology for Agent Societies

In this chapter, we present a framework for the design of agent societies that considers the influence of social organizational aspects on the functionality and objectives of the agent society and specifies the development steps for the design and development of an agent-based system for a particular domain. Our approach will provide a generic frame that directly relates to the organizational perception of the problem. The framework specifies the development steps of the design and development of an agent-based system for a particular domain. Based on the coordination characteristics of a domain, the methodology provides three frameworks for societies (market, hierarchy, and network). These frameworks relate to the organizational perception of a problem and allows for existing methodologies to be used for the development, modeling, and formalization of each step. The methodology supports the development of increasingly detailed models of the society and its components.

The Logic Behind Negotiation

The use of agents in Electronic Commerce environments leads to the necessity to introduce some formal analysis and definitions. A four-step method is introduced for developing EC-directed agents, which are able to take into account nonlinearites such as gratitude and agreement. Negotiations that take into account a multistep exchange of arguments provide extra information, at each step, for the intervening agents, enabling them to react accordingly. This argument-based negotiation among agents has much to gain from the use of Extended Logic Programming mechanisms. Incomplete information is common in EC scenarios; therefore, arguments must also ta

A Formal Method for the Development of Agent-Based Systems

Recent advances in testing and verification of software based on formal specifications of the system to be built have reached a point where the ideas can be applied in a powerful way in the design of agent-based systems. The software engineering research has highlighted a number of important issues: the importance of the type of modeling technique used; the careful design of the model to enable powerful testing techniques to be used; the automated verification of the behavioral properties of the system; and the need to provide a mechanism for translating the formal models into executable software in a simple and transparent way. This chapter presents a detailed and comprehensive account of the ways in which some modern software engineering research can be applied to the construction of effective and reliable agent-based software systems. More specifically, we intend to show how simple agents motivated from biology can be modeled as X-machines. Such modeling will facilitate verification and testing of an agent model, because appropriate strategies for model checking and testing are already developed around the X-machine method. In addition, modular construction of agent models is feasible, because X-machines are provided with communicating features, which allow simple models to interact.

Cooperation Between Agents to Evolve Complete Programs

In this paper, we present a multiagent system approach with the purpose of building computer programs. Each agent in the multiagent system will be in charge of evolving a part of the program, which in this case, can be the main body of the program or one of its different subroutines. There are two kinds of agents: the manager agent and the genetic programming (GP) agents. The former is in charge of starting the system and returning the results to the user. The GP agents include skills for evolving computer programs, based on the genetic programming paradigm. There are two sorts of GP agents: some can evolve the main body of the program and the others evolve its subroutines. Both kinds of agents cooperate by telling each other their best results found so far, so that the search for a good computer program is made more efficient. In this paper, this multiagent approach is presented and tested empirically.

SoFAR

SoFAR, the Southampton Framework for Agent Research, is a versatile multiagent framework designed for Distributed Information Management tasks. SoFAR embraces the notion of proactivity as the opportunistic reuse of the services provided by other agents, and it provides the means to enable agents to locate suitable service providers. The contribution of SoFAR is to combine ideas from the distributed computing community with the performative-based communications used in other agent systems: communications in SoFAR are based on the startpoint/endpoint paradigm, a powerful abstraction that can be mapped onto multiple communication layers. SoFAR also adopts an XML-based declarative approach for specifying ontologies and agents, providing a clear separation with their implementation.

Machine Learning for Agents and Multi-Agent Systems

In order to be truly autonomous, agents need the ability to learn from and adapt to the environment and other agents. This chapter introduces key concepts of machine learning and how they apply to agent and multi-agent systems. Rather than present a comprehensive survey, we discuss a number of issues that we believe are important in the design of learning agents and multi-agent systems. Specifically, we focus on the challenges involved in adapting (originally disembodied) machine learning techniques to situated agents, the relationship between learning and communication, learning to collaborate and compete, learning of roles, evolution and natural selection, and distributed learning. In the second part of the chapter, we focus on some practicalities and present two case studies.

Architectures for Cognitive and A-Life Agents

In this chapter, research into the nature of drives and motivations in computational agents is visited from a perspective drawing on artificial life and cognitive science. The background to this research is summarized in terms of the possibility of developing artificial minds. A particular cognitive architecture is described in terms of control states. Steps toward producing an implementation of this architecture are described by means of experimentation into the nature of specific control states. The design of a simple a-life architecture for a predator–prey scenario is described using a transition state diagram. This architecture is then used as a platform with which to develop an agent with drives. This second architecture is used to develop an agent with explicit motivations. The discussion of these (and other) experiments shows how these simple architectures can help to provide some answers to difficult research questions in cognitive science.