In this paper, the implications of applying the idea of gift exchange mechanism, inspired from Pierre Bourdieu's sociological theories, into a market-based multiagent system are explored. Our work is directed in the continuation of investigations by Knabe (2002), who addressed the formation of different organizations structures between providers in a profit-oriented market. We nevertheless scrutinize various hypotheses centered to gift exchange in which an agent sacrifices its profit for a long-term binding relationship. The idea is to aim a larger profit through alliances that are formed as an effect of gift exchange. Our suggestion is that a multiagent system (MAS) based on the social mechanism of gift exchange performs a high level of robustness and durability. The market in our case comprises of customers and providers agents. The former calls for proposals for the tasks they introduce in the market, while the latter proceed with the execution of tasks based on their abilities and other circumstances. In well defined cases, the providers are able to delegate tasks to other providers. This allows them to give presents to other providers so that the gift exchange mechanism becomes possible. The agents are either profit-oriented or the ones who prefer exchanging gifts and are in pursuit of others who also practice this mechanism. A number of interesting scenarios are examined that include preservation of a hierarchical structure in the market, situations resulting in the forming of an alliance between two providers, and split of profit-oriented and gift-giving agents.

In sociology and distributed artificial intelligence, researchers are investigating two different ways of scaling. On the one hand, there is qualitative scaling, meaning that (social) complexity is increased, introducing regular practices of action, institutions, new fields of social action and requiring new dimensions in perception and decision making. On the other hand, researchers are interested in investigating quantitative scalability, i.e. how goals can be achieved under the constraints imposed by a growing population. Our argument is structured as follows: firstly, we want to establish that organizations and interorganizational networks are an important cornerstone for the analysis of qualitative scaling. Secondly, we show by empirical evaluation that an elaborate theoretical concept of such networks increases the quantitative scalability of multiagent systems.

We propose FORM, a new Framework for self-Organization and Robustness in Multiagent systems. This framework supports the design of task-assignment multiagent systems in a way that is informed by sociological theory. It is founded on the habitus-field-theory of sociologist Pierre Bourdieu. In accordance to this theory, we consider the special quality of "organization" as an autonomous and self-organizing social entity with clear distinction to the coordination via social interactions. Organizations are viewed as both "autonomous social fields" and "corporate agents", which are competing with other organizations in the same domain. While our framework makes no claims about an underlying agent architecture, it consists of a matrix of mechanisms for delegation (task delegation and social delegation), which we consider a central concept to define organizational relationships. Using this matrix as a basic toolset, we propose a spectrum of seven types for the structure of multiagent systems, defined by qualitatively different relationships.

This thesis reports on work conducted under the Schwerpunktprogramm Sozionik,1 a basic research project funded by the Deutsche Forschungsgemeinschaft to transfer knowledge between sociology and distributed artificial intelligence (DAI). The motivation of this work is to use sociological notions and theories of ’organisation’ to blueprint more robust multiagent systems. Based on an analysis of the DAI literature, we give a precise and empirically verifiable definition of robustness, which we call τ-robustness. This notion consists of (i) the definition of a performance measure, (ii) the definition of a perturbation against which the performance measure is evaluated, and (iii) a threshold τ , which marks the maximally allowed deviation of the performance measure to call the tested system τ-robust. The main theoretic contribution of this thesis is a framework of design parameters for robust multiagent organisation based on a sociological notion of organisation. This framework, called the Framework for selfOrganisation and Robustness in Multiagent systems (FORM), can be applied to multiagent organisation using the concept of holonic multiagent systems, for which a new, extended definition is presented. By freely combining all possible values of the parameters in the framework, it allows to model more than 90 000 different forms of multiagent organisation. Furthermore, we discuss how the holonic design parameters relate to different dimensions of the autonomy of the participating agents. This discussion shows that, and how, the notion of self-organisation is connected to the notion of adjustable autonomy. As the notion of autonomy is central to the definition of an agent, this constitutes an important contribution to the general theory of multiagent systems. In order to make the large design space spanned by FORM concrete, we chose to model a subset of the possible organisational forms. This choice is based on a sociological analysis of organisations in today’s economy and makes a diverse use of the available values for the design parameters. The modelled forms of organisation are arranged on a spectrum of autonomy ranging from fully autonomous agents to more and more coupled agents, until we reach an organisational form where the boundaries between individual agents dissolve and only one single agent remains. This spectrum can be used to devise a mechanism for self-organisation in the sense that agents start to organise in a loosely coupled holon and increase the 1 Collaborative research group in Socionics.

Starting from a general definition of how to model the organisation of multiagent systems with the aid of holonic structures, we discuss design parameters for such structures. These design parameters can be used to model a wide range of different organisational types. The focus of this contribution is to link these design parameters with a taxonomy of different types of autonomy relevant in multiagent organisation. We also discuss the constraining effect of autonomy on the recursive nesting of multiagent organisation. As the domain for applying multiagent systems we choose a general view on multiagent task-assignment.

Market-based approaches have a good tradition for supporting the development of task-assignment multiagent systems. Such systems consist of customer agents that have jobs to assign, and provider agents that have the resources to perform the jobs. Jobs can be complex, requiring the collaboration of several provider agents. We present a set of sociological forms of collaboration between firms that have the potential to increase performance through the structure they impose. This gain of structure, which comes with a loss of autonomy, is especially valuable in settings where communication is limited, which is an appropriate assumption in large-scale applications. We empirically evaluate these organizational forms according to the amount of communication required and the rate of failed task-assignments. Furthermore, we investigate the behaviour of each form in the face of agents dropping out during runtime and compare them to settings without organizational forms.

With the growing usage of the world-wide ICT networks, agent technologies and multiagent systems are attracting more and more attention, as they perform well in environments that are not necessarily well-structured and benevolent. Looking at the problem solving capacity of multiagent systems, emergent system behaviour is one of the most interesting phenomena, however, there is more to multiagent systems design than the interaction between a number of agents: For an effective system behaviour we need structure and organisation. But the organisation of a multiagent systems is difficult to specify at design time in the face of a changing environment. This paper presents basic concepts for a theory of holonic multiagent systems to both provide a methodology for the recursive modelling of agent groups, and allow for dynamic reorganisation during runtime.

With the growing amount of internet users, a negative form of sending email spreads that affects more and more users of email accounts: Spamming. Spamming means that the electronic mailbox is congested with unwanted advertising or personal email. Sorting out this email costs the user time and money. This paper introduces a distributed spam filter, which combines an off-the-shelf text classification with multiagent systems. Both the text classification as well as the multiagent platform are implemented in Java. The content of the emails is analyzed by the classification algorithm 'support vector machines'. Information about spam is exchanged between the agents through the network. Identification numbers for emails which where identified as spam are generated and forwarded to all other agents connected to the network. These numbers allow agents to identify incoming spam email. In this way, the quality of the filter increases continuously.

Market-based approaches have a long tradition in supporting of task-assignment multiagent systems. Such systems consist of customer agents with jobs to assign, and provider agents that have the resources to perform these jobs. Jobs can be complex in the sense that they require the collaboration of several provider agents. We present a set of organisational forms of collaboration between firms that have the potential to increase performance through the structure they impose. This gain of structure, which comes with a loss of autonomy of the individual agents, is especially valuable in settings where communication has to be limited.

With the growing usage of the world-wide information technology networks, agent technologies and multiagent systems are attracting more and more attention, as they perform well in environments that are not necessarily well-structured and benevolent. Looking at the problem solving capacity of multiagent systems, emergent system behaviour is one of the most interesting phenomena. But there is more to multiagent system design than the interaction between a number of agents: For effective system behaviour we need structure and organisation. Moreover, it is difficult to specify the organisation of a multiagent systems in a changing environment at design time. The theory of holonic multiagent systems promises both, to provide a methodology for the recursive modelling of agent groups and to allow for dynamic reorganisation during runtime.

In this paper we introduce a new approach, the so-called AgentComponent (AC) approach which combines component and agent technology. A multi agent system (MAS) is composed of AC instances, each AC instance consists of a knowledge base, storing the beliefs of an AC instance, of slots, storing the communication partners of an AC instance, of a set of ontologies, that represent domain specific languages for certain contexts, and of so-called ProcessComponents (PC) representing the behaviours of an AC instance. The AC is a generic component that can be reused (instantiated ACs) and parametrized by customizing the communication partners (slots), the ontologies and the behaviours (PCs) that can be added and removed from any AC instance. Hereby we achieve added value for agents and components. Agents can be easily composed, customized and reused whereas components get enhanced communication and interaction facilities from agents. We present this approach in detail, show how to construct a component-based MAS by a simple example and present a graphical tool for composing systems of AgentComponents.

Market-based approaches for task-assignment multiagent systems consist of customer agents with jobs to assign, and provider agents that have the resources to perform these jobs. Jobs can be complex in the sense that they require the collaboration of several provider agents. We present a set of sociological forms of inter-organizational networks that have the potential to increase performance through the structure they impose on collaboration.This gain of structure is especially valuable in settings where communication is limited, which is an appropriate assumption in large-scale applications. We empirically evaluate these organizational forms according to the amount of communication required and the rate of failed task-assignments, and compare them to a system without organizational forms. Furthermore,we investigate the effect of letting agents choose at runtime in which kind of organizational form to engage and which other agents to choose for this collaboration.Our evaluation shows that the proposed organizational forms and mechanisms for self-organization have the ability to improve the efficiency of a market-based multiagent system.

Engineering Agent-Based Systems.- The AgentComponent Approach, Combining Agents, and Components.- From Simulated to Real Environments: How to Use SeSAm for Software Development.- Indicators for Self-Diagnosis: Communication-Based Performance Measures.- Systems and Applications (1).- The AEP Toolkit for Agent Design and Simulation.- On Programming Information Agent Systems - An Integrated Hotel Reservation Service as Case Study.- Applying Agents for Engineering of Industrial Automation Systems.- Systems and Applications (2).- SimMarket: Multiagent-Based Customer Simulation and Decision Support for Category Management.- A Multi-agent Approach to the Design of an E-medicine System.- Implementing Heterogeneous Agents in Dynamic Environments, a Case Study in RoboCupRescue.- Models and Architectures.- Model for Simultaneous Actions in Situated Multi-agent Systems.- Handling Sequences of Belief Change in a Multi-agent Context.- From the Specification of Multiagent Systems by Statecharts to Their Formal Analysis by Model Checking: Towards Safety-Critical Applications.- The Semantic Web and Issues of Inter-operability.- The SWAP Data and Metadata Model for Semantics-Based Peer-to-Peer Systems.- An Ontology for Production Control of Semiconductor Manufacturing Processes.- Ontology-Based Capability Management for Distributed Problem Solving in the Manufacturing Domain.- Using the Publish-Subscribe Communication Genre for Mobile Agents.- Issues of Collaboration and Negotiation.- Multiagent Matching Algorithms with and without Coach.- Improving Evolutionary Learning of Cooperative Behavior by Including Accountability of Strategy Components.- The C-IPS Agent Architecture for Modeling Negotiating Social Agents.

Multiagent systems (MAS) have found their way into industrial applications in recent years and appear to be one of the most promising technologies that originated in AI research in recent years. However, evaluation standards as they are common e.g. in the scheduling or database systems communities are largely amiss. In this paper, we propose communication-based performance measurement (CBPM) as a new method that is particularly suitable for open, communication-intensive MAS, and argue that it can be used as to design indicators for self-diagnosis by the MAS itself. The ability of such self-diagnosis is a prerequisite for MAS with self-repairing and self-optimising properties required by the autonomic computing view. CBPM is based on the idea that important aspects of the external behaviour of a MAS can be measured in terms of the communication processes within them. We present different levels of communication-based performance measurement: frequency analysis of performatives and analysis of complex message patterns. Several examples of analyses of inter-agent communication based on FIPA-ACL and the contract-net protocol in implemented, complex, market-oriented MAS demonstrate the usefulness of our approach. We conclude that these performance measures provide useful information about MAS and pave the way for devising autonomic self-improvement methods for these systems.

In this paper we suggest a new sociological concept to the study of (self-) organization in multiagent systems. First, we discuss concepts of (self-) organization typically used in DAI. From a sociological point of view all these concepts are missing the special quality of organizations as self-organizing social entities. Therefore we present a concept of organization based on the habitus-field theory of Pierre Bourdieu. With reference to this theory, organizations are viewed as both “autonomous social fields” and “corporate agents” which are competing with other organizations in the same domain. Finally, we describe the Framework for Self-Organization and Robustness in Multiagent systems (FORM) corresponding to these sociological characteristics of organizations. This framework uses delegation as the central concept to define organizational forms and relationships in task assignment multiagent systems.

No matter if a population is human or artificial, we can surely identify phenomena that can be described as micro or macro phenomena. In this paper, we discuss micro and macro aspects of a population from a DAI and a sociological point of view. We analyse similarities and differences in these viewpoints, and identify misperceptions in the DAI community about the micro-macro terminology. We explain these misperceptions and argue for the transfer of sociologically founded concepts to agent-based social simulation. Our research is done in the DFG focus programme socionics. We cooperate with sociologists from University Hamburg-Harburg with the intention to transfer knowledge from sociology to DAI as well as from DAI to sociology. In cooperation with DFKI Saarbrü = cken we work on improving agent theories to be applied in large sized multi-agent systems in the freight logistics domain.