Agreement Technologies

In this chapter we analyse the concept of agreement from a Computing perspective. In particular, we argue that the capability of software components to dynamically forge and execute agreements at run-time will become increasingly important, and identify key areas and challenges that need to be addressed in order to advance in this direction. Finally, we introduce the emerging field of Agreement Technologies for the construction of large-scale open distributed software systems, and identify technologies that are in the sandbox to define, specify and verify such systems.

Relationships between empirical and philosophical approaches to the law have not been always peaceful. Agreement seems the most natural way to build up and implement regulations and justice within human-machine interfaces (natural and artificial societies), and might help to bridge the gap between both theoretical approaches. Recent researches on

relational law

,

relational justice

,

crowdsourcing

,

regulatory systems and regulatory models

are introduced. These concepts need further clarification, but they stand as political companions to more standard conceptions of law in the Semantic Web.

Building on the assumption that agreements are primarily useful for optimizing social interactions in view of individual interests (greasing society) rather than being meant to keep us together (gluing society), I offer a bird’s-eye view of several topics of interest in agreement theory for cognitive and social science. This brief review is far from comprehensive, and focuses instead on specific themes: a socio-cognitive analysis of the distinction between “being in agreement” and “having an agreement”, the multiple paths leading to agreement (e.g., norms, organizations, argumentation), how these phenomena both presuppose and facilitate agreements, and why a certain level of disagreement is to be desired even in the most well-ordered social system. The relevance of these considerations for the development of agreement technologies will also be explored.

Here we discuss the relationship between Agreement Technologies and the Semantic Web, especially focusing on how Semantic Web standards play a role in the Agreement Technologies stack, but also issues related to Linked Data and the Web of Data. We start the chapter with an account of Semantic Web standards and ensue with the role they play in Agreement Technologies.

This chapter provides an overview of the logical formalisms that have been proposed to define the formal semantics of knowledge systems that are distributed, heterogeneous and multi contextual. It starts with the abstract notions that are common to many of these logics, and then focuses on individual formalisms.

In open, dynamic and distributed systems, it is unrealistic to assume that autonomous agents or peers are committed to a common way of expressing their knowledge, in terms of one or more ontologies modelling the domain of interest. Thus, before any kind of communication or cooperation, agents must reach an agreement on the meaning of the terms they use for structuring information, conceptualizing the world, or representing distinct entities. Over the years several approaches have been proposed for semantic agreement driven by ontology matching in a distributed setting: argumentation-based models, constraint satisfaction methods and probabilistic models. The aim of this chapter is to present a brief overview of the state-of-the-art on these approaches and discuss the main open issues and challenges for future research. We firstly introduce the ontology matching process for semantic agreements and the notion of argumentation frameworks, and then we present scenarios applying such frameworks. Next, we specify the problem of synthesizing different matching methods as a constraint optimization problem and show the benefits of this approach and we present an approach for peers organized in arbitrary networks to reach semantic agreement on their correspondences. Finally, we discuss some open issues and future research directions for semantic agreement based on ontology matching.

In this chapter we discuss how

semantic technologies

in general and specific

Semantic Web standards

in particular can contribute to the goal of achieving

interoperability

between independent, loosely coupled, heterogeneous, autonomous software components (i.e. agents) and for the realization of open interaction systems. In particular we will discuss how those technologies have been used for the definition of the semantics of agent communication languages, for the definition of norms and policies used to regulate interactions in open frameworks, and for defining efficient mechanisms for matching demands (i.e., content they need) to supplies (i.e., available content) in telecommunication networks. In particular regarding this last type of application we describe a techno-economic approach for solving the matching problem, by means of a multi-agent system representing an electronic marketplace. Its functionality is realized by applying a semantic-aware content discovery model with two-level filtering in order to finally recommend a ranked set of eligible content to the users in response to their requests. The filtering processes not only consider the semantic information associated with the available content, but also ratings regarding the actual performance of businesses that act as content providers as well as the prices paid by businesses for advertising their content.

This chapter introduces Web Service discovery in heterogeneous environments. While most contemporary service discovery systems assume that both service request and advertisement are represented in the same description model, this assumption does not always hold in a heterogeneous environment where different types of agents interact. An agreement on a common model for both parties involved in the communication is required. This chapter presents three approaches to dealing with service description model mismatch: (1) translation among existing models; (2) neutral model built on the intersection of existing models; (3) neutral model built on the union of existing model. Representation and utilization of the domain ontology alignment results for semantic annotated service description are discussed in the second part of this chapter.

This chapter shows how semantic technologies can enable the development of an agent-based intelligent middleware for the Grid (the

Agents in Grid

system). Resource providing agents are organized in teams and negotiate contracts with agents representing users. All information is ontologically demarcated and semantically processed. In particular, user preferences are turned into ontology class expressions and are applied directly in contract negotiations. The chapter provides a detailed overview of how ontologies are used in the

AiG

system. Key issues in the development of the

AiG

ontology are discussed, and the implementation of the ontology-focused parts of the system are described.

Deontic logic is the logic of obligation and permission. In the literature it has mainly been studied in terms of a list of problems and that is the way we chose to present it in this section. There are three main categories of problems. The first category is concerned with the nature of norms. The second category concerns phenomena of conflict, violation and revision. Finally, the third category studies deontic phenomena in the context of other logical structures.

This chapter aims to identify the main relevant steps in the evolution of (social) norms as well as some of the factors or determinants of such a process and to discuss the most urgent scientific tasks to be fulfilled within a community of scientists committed to the study of norms. It is clearly the case that the scientific study of norms needs innovation and opening up to new instruments, new tools, new competencies, and especially new perspectives and approaches. As argued in this chapter, the merging of Agent-Based Modelling and Multi-Agent Systems appears as a promising direction to give a strong, innovative boost to the study of norms.

This chapter summarizes two main views on norms and games as emerging in literature on game theory, social science, philosophy and artificial intelligence. The first view originates in the field of mechanism design or implementation theory and characterizes norms as mechanisms enforcing desirable social properties in classes of games. According to the second view, originating from work in the social sciences and evolutionary game theory, norms are studied via the notion of equilibrium and are viewed as emergent social contracts or conventions.

This chapter provides an overview of the main lines of inquiry of Artificial Intelligence and Law (AI and Law), such as rule-based reasoning, case-based reasoning, ontologies, argumentation, theory construction and legal deontics. We argue that a rich picture of the law is emerging from the AI and Law research, which can complement and integrate not only research in law and legal theory, but also other attempts to provide formal and computational models of norms.

In open dynamic societies, agents are required to work with others that do not necessarily have the same set of objectives. If left unchecked, self-interested agents will try to accomplish their individual goals without regard for others. Norms provide a means to regulate agent behaviour, and this requires some consideration of the ways in which norms impact on agent reasoning and behaviour. In this chapter, we review work on the development of normative agents, focussing on agent architectures in which action is determined by norms in a system or environment.

In this chapter, some different (and in part complementary) analyses and approaches to the study of the relationship between Norms and Trust are briefly introduced. First, how Trust and Norms can be considered the basis of each other is analyzed depending on the phenomenon considered. Second, starting from the fact that an agent’s trustworthiness can be evaluated on its compliance with norms, the authors consider the different ways to comply with a norm and the relationships with this analysis and the trust models. In particular the feedback on the norm adaptation. Finally, an interesting analogy between Obligation, Role and Information Scenarios – with respect to the intimate connection between trust and rules – is presented. The need to consider the volitional component in the trust concept and the corresponding link between some types of emotions (regret, anxiety, hope) is also presented.

Norms and argumentation are two research areas which have been becoming increasingly connected over the last decade, in the legal field, in knowledge representation, ethics, or linguistics, and most recently, in agreement technologies in computer science. Norms are used to set the space of legal agreements (or commitments) and argumentation is used to choose among the possible agreements. Moreover, we may consider norms set not only the scope of possible legal agreements, but also the way we can choose among these possible agreements. We highlight the main open challenges in argumentation and norms considering two parallel problems: how to argue about norms, and how norms influence the argumentation.

This chapter addresses how agent organisations can improve and accelerate coordination processes in open environments. A state-of-art of recent proposals for describing agent organisations is given, relating the different methodologies and formal approaches for defining agent organisations in an explicit way. As example, four different proposals developed within the COST action IC0801 are detailed: (i) the MOISE organisation Model, which provides structural, functional and normative specifications of an organisation, and it is integrated in an Organisation Management infrastructure; (ii) the Virtual Organisation Model, which describes the structural, functional, dynamical, environmental and normative dimensions of an organisation, and it is complemented by the Virtual Organisation Formalization; (iii) the Agent-Oriented Modelling for sociotechnical systems, which are organisations consisting of human and man-made agents; and (iv) the AAOL agent architecture, in which groups of autonomous agents are organized in Localities. This chapter proposes a global comparison of different existing organisational models with the four detailed models in terms of the different description dimensions they propose.

Everyday uses of the notion of institution and some typical institutions have been studied and formalized by economists and philosophers. Borrowing from these everyday understandings, and influenced by their formalizations, the notion of institution has been used within the agents community to model and implement a variety of socio-technical systems. Their main purpose is to

enable

and

regulate

the interaction among autonomous agents in order to achieve some collective endeavour. In this chapter we present and compare three frameworks for agent-based institutions (i) ANTE, a model that considers electronic institutions as computational realizations of adaptive artificial environments for governing multi-agent interactions; (ii) OCeAN, extended in MANET, a model for specifying Artificial Institutions (AIs), situated in agent environments, which can be used in the design and implementation of different open interaction systems; and (iii) a conceptual core model for Electronic Institutions (EIs), extended with EIDE, based on open, social, decomposable and dialogical interactions. Open challenges in the specifications and use of institutions for the realization of real open interaction systems are discussed.

From an agent architecture perspective and analysing the reasoning capabilities of agents with respect to organisations, different cases may be considered: agents may or may not have an explicit representation of the organisation, and they may or may not be able to reason about it. Organisational reasoning agents have the capability to represent the organisation and are able about reason on it. In this chapter, we will discuss the main features of this kind of agents and which are the fundamental mechanisms for reasoning about organisations. We will also describe some approaches proposed in the literature related to how agents can take decisions on their participation in an organisation.

The capability of autonomous adaptation to changing conditions is a feature that requires the ability of agents to alter their own configuration and even their own composition. Adaptive agent organisation should consider agent self-adaptation and reorganisation processes. In this chapter a review of methods for designing and/or implementing adaptive agent organisations is given.

We discuss the value of argumentation in reaching agreements, based on its capability for dealing with conflicts and uncertainty. Logic-based models of argumentation have recently emerged as a key topic within Artificial Intelligence. Key reasons for the success of these models is that they are akin to human models of reasoning and debate, and their generalisation to frameworks for modelling dialogues. They therefore have the potential for bridging between human and machine reasoning in the presence of uncertainty and conflict. We provide an overview of a number of examples that bear witness to this potential, and that illustrate the added value of argumentation. These examples amount to methods and techniques for argumentation to aid machine reasoning (e.g. in the form of machine learning and belief functions) on the one hand and methods and techniques for argumentation to aid human reasoning (e.g. for various forms of decision making and deliberation and for the Web) on the other. We also identify a number of open challenges if this potential is to be realised, and in particular the need for benchmark libraries.

Automated negotiation provides an important mechanism to reach agreements among distributed decision makers. It has been extensively studied from the perspective of e-commerce, though it can be seen from a more general perspective as a paradigm to solve coordination and cooperation problems in complex systems, e.g., task allocation, resource sharing, or surplus division. A variety of negotiation models have been proposed according to the many different parameters which may characterize a negotiation scenario. In this chapter, we briefly review the key concepts about multi-attribute negotiation and the most relevant works in the field, and then we focus on one of the more challenging topics on the field in the last few years, namely complex negotiations. In particular, we focus on situations where unanimous agreement is not possible or simply not desired, which is very common in negotiations involving complex, non-monotonic utility spaces. We describe a framework with which to perform multiagent negotiations where we can specify the type of agreements needed in terms of utility sharing among the agents. The proposed multi-round mediation process is based on the analysis of the agents’ offers at each negotiation round. At each round, the mediator applies Global Pattern Search (GPS) to the offers and then a linguistic expressed mediation rule based on Ordered Weighted Averaging Operators (OWA) that formalizes the consensus policy. At each round this mediation process generates a social contract that is submitted as feedback to the agents.

Trust and reputation are two distinct social constructs of high complexity that have been studied for decades in different areas of knowledge. In order to allow for efficient models of computational trust and reputation, one must first understand the nature and dynamics of each one of these constructs. In this chapter, we focus on the social and cognitive aspects of the trust concept, and overview its fundamental characteristics, such as its determinants, nature, and dynamics. Then, we present two distinct hypothesis one can state for the interplay between trust and reputation: either reputation is an antecedent of trust, or both are considered as two distinct contributions to the ultimate decision making process. If they are seen as isolated components, trust is no longer directly influenced by reputation. Finally, we briefly refer to current existing computational trust models, including those that integrate the management of computational reputation.

This chapter presents an anthropocentric method for trust management in the area of agreement technologies, called Qualitative Assessment Dynamics. As opposed to other approaches like Bayesian inference, game-theoretic approaches and alike, this method focuses on linguistic bases for modeling trust in agents societies. Therefore it deploys operands and operators that mimick trust dynamics by reflecting related mental processes through corresponding descriptions in ordinary languages.

In this chapter we discuss the ways in which trust can be combined with argumentation. This is a new field of research that is showing promising approaches to a number of problems in both argumentation and trust. We discuss three ways in which trust and argumentation are combined. The first is to use the trustworthiness of an agent as a level of confidence in the arguments it provides. Parsons et al. take this a step further and consider the ramifications for this in the combination of arguments from different sources. The second way that trust and argumentation are combined is to compute the trustworthiness of an agent based on arguments about its behavior and we discuss two different approaches to this. Finally, argumentation can be used to improve communication about trust. Methods for doing this are discussed at the end of this chapter.

In this chapter we discuss the problem of communicating about trust and how semantic technologies can help. We briefly introduce these semantic technologies and then discuss two well-known ontologies of trust:

$${\mathcal{L}}_{Rep}$$

and FORe. However, defining a shared language for trust ignores the personal and subjective aspect of trust, which are an important part of how it is used. We therefore discuss a number of filtering and alignment methods that allow for the processing of communicated trust evaluations without compromising the subjective aspect of trust.

Most of the current reputation models suffer from some common vulnerabilities. This means that malicious agents may be able to perform attacks that exploit these vulnerabilities, which has the potential to cause much harm such as monetary losses and to place the whole system in jeopardy. In this chapter we detail some of the most important vulnerabilities of current reputation models. We also detail examples of attacks that take advantage of these vulnerabilities in order to achieve strategic manipulation of reputation models. Moreover, we review works that partially/fully address these vulnerabilities, and thus, prevent possible attacks from being successful.

An important role of reputation in the context of organisations is to implement mechanisms for norm enforcement and incentives so that agents behave as expected by the organisation they belong to. However, an organisation also provides other useful elements to compute reputation for that issue. As a collective phenomenon, reputation is better thought of as embodied in an organisation. The chapter “Reputation and Organisations” discusses the links and opportunities raised by the integration of a multiagent reputation system within organisations. A first study is dedicated to the organisational concepts and entities that may contribute to reputation mechanisms. A survey of the elements of an organisation that can either be considered as trustees for social evaluation or act as informational sources is provided. The chapter describes then how reputation mechanisms can be seen and implemented as organisational processes. Finally, a few future research challenges of this combined view are proposed.

In this chapter we put together two major threads of work: trust in the enactment of contracts and the modelling of relationships between agents. We depart from previous work where trust is defined as the relationship between commitment of action and instant observation of the actual actions being performed. Here we generalise the approach by assuming a time delay between the observation of the actions and their valuation. The fundamental new idea being that commitment for action has a social dimension as the commitment of an agent should mean ‘attempting to act’ in the interest of the contractual partner, and that attempt has a time dimension that cannot be ignored.

In this chapter, we present a case-based argumentation system that allows the technicians of a call centre to reach agreements and provide high quality customer support. Our application implements an argumentation framework that proposes software agents that represent technicians and have individual knowledge resources to automatically generate their positions and arguments. On the one hand, agents have a domain-cases case-base, with domain-cases that represent previous problems and their solutions. On the other hand, agents have an argument-cases case-base, with argument-cases that represent previous argumentation experiences and their final outcomes. These resources are represented by using ontologies. In this research, previous work that deployed a case-based multi-agent system in a real call centre has been enhanced by allowing agents to argue about the best way of solving the incidents that the call centre receives. The infrastructure that gives support and implements these argumentative agents has been developed and tested. The results of these tests are shown in this chapter.

The use of agreement technologies in the planning and execution of goods transports is analyzed. We have previously suggested an approach called Plug and Play Transport Chain Management (PnP TCM) that provides agent-based support for key tasks, such as, finding the best sequence of transport services for a particular goods transport, monitoring the execution of the transport, and managing the interaction between the involved actors. In this paper we analyze five agreement technologies in the context of PnP TCM, i.e., semantics, norms, organizations, argumentation and negotiation, and trust. We conclude that all five technologies play a critical role in the realization of PnP TCM.

The ANTE framework encompasses results of research efforts on three main agreement technology concepts, namely

negotiation

,

normative environments

and

computational trust

. ANTE has been conceived as a general framework with a wide range of applications in mind. This chapter provides an overview of the main guidelines of this project, and explores two application domains for this framework: automated B2B electronic contracting, and disruption management in the context of an airline company operational control.

We propose an electronic institution approach to build a virtual market as an open multi-agent system that handles several negotiation protocols in a coherent and flexible fashion. This proposal has been inspired by the work in

mWater

, which is developed as a regulated environment where autonomous agents trade rights for the use of water in a closed basin. We also present a generic negotiation framework that is enabled with tools to specify performance indicators, to spawn agent populations and allow humans, as well as software agents, to participate in simulations of water-right virtual trading. This demonstrates an interesting aid for data organisation, and for communication and negotiation among the different stakeholders of a basin.

Nowadays, governments are increasingly taking advantage of Information and Communication Technologies to provide services over the internet (the so-called e-Government applications) to citizens, businesses, employees, and agencies. We argue that e-government services will benefit from being distributed and intelligent, and thus, that they can be modelled as Multi-Agent Systems (MAS). The field of MAS focuses on the design and development of systems composed of autonomous entities which interact within an environment in order to achieve their common or individual goals. Nevertheless, although humans can be seen as autonomous entities, most MAS methodologies and infrastructures do not consider direct human participation. In general, the human role is limited to acting behind the scenes by customising provided agent templates. The resulting agents participate in the system on humans’ behalf. In order to overcome this limitation we propose using 3D Virtual Worlds, which is one of a very few technologies that provides all the necessary means for direct human inclusion inside software systems. 3D Virtual Worlds are 3D graphical environments where humans participate represented as graphical embodied characters (avatars) and can interact there by using simple and intuitive control facilities. We advocate that 3D Virtual Worlds technology can be successfully used for “opening” multiagent systems to humans. This idea is used in Virtual Institutions, which combine Electronic Institutions and 3D Virtual Worlds to engineer applications where participants may be human and software agents. In this chapter we present the prototype

v-mWater

, a

v

irtual

m

arket based on trading

Water

. It is an e-Government application in the agriculture domain modelled as a Virtual Institution where participants are irrigators and employees of a hydrographic basin. We present the specification of the system, the virtual world generation from this specification and its deployment using the Virtual Institution eXEcution Environment (VIXEE). Finally, we show an example execution of our virtual market in order to illustrate the advantages of our approach.

In this Chapter we propose an organization-based multi-agent application for emergency medical assistance (EMA). The application uses different

Agreement Technologies

(AT) to provide support to the entire process of out-of-hospital assistance to severe emergency patients and to all involved participants. The system is inspired by the operational model of the Emergency Medical Coordination Centre of the Autonomous Region of Madrid in Spain:

SUMMA112

. The application is also intended to reduce the average travel times of ambulances to emergency patients by making efficient use of the available resources. Regarding the latter, three different coordination mechanisms are proposed to optimize the allocation of ambulances to patients: one based on trust, an auction-based negotiation model, and auction-based negotiation with trust. We test these mechanisms in different experiments. The results empirically confirm that using AT based mechanisms can reduce the average response times of EMA services.

This chapter describes an environment to support the rapid assembly of agent-oriented business collaborations. Our environment allows: (i) setting up a collaboration environment as a virtual organization; (ii) reaching agreements within the collaboration environment to form short-term business collaborations; (iii) enacting business collaborations; and (iv) tracking the performance of agents within business collaborations to build their trust and reputation within the collaboration environment.

In this paper, we claim that the online selling process can be improved if the experience of the customer is closer to that of one in a retailing store. For this purpose, we provide a virtual selling agent that is persuasive and adaptive. Our persuasive selling agent initiates the dialogue and drives it in order to simultaneously advance two processes: customer profiling and the identification of relevant proposals which are supported by arguments. Moreover, our virtual seller is adaptive since she is able to adjust her behaviour and her arguments according to the customer profile.