Animated specifications of computational societies

Abstract

E-markets and negotiation protocols are two types of application domains that can be viewed as open computational societies. Key characteristics of such societies are agent heterogeneity, conflicting individual goals and limited trust. The risk that members of such societies will not conform to specifications imposes the need for a framework that will facilitate the designers to determine to what extent it is desirable to deploy their agents in such societies. We address this need by presenting a formal framework for specifying, animating, and ultimately reasoning about and verifying the properties of open computational systems. We view computational systems from an external perspective, aiming to account for the institutional and social aspects of these systems. We identify the key concepts and illustrate how they are used by formalising an example employing the contract net protocol. The framework and associated logical inferences have been implemented as a software platform that provides automated animation of the global states of an open system (society) during its execution. Simulations have demonstrated that the implementation of the framework establishes a foundation for a rich, formal representation of open computational societies.

Automated Analysis of Auction Traces

“…The verification of such properties upon auctioning traces has however not been addressed before. In [2] a framework is introduced for the specification of properties for open systems, as well as reasoning and verification of these properties. They use the contract net protocol as a running example.…”

Peer Pressure as a Driver of Adaptation in Agent Societies

“…-open: in the sense of Artikis et al [1] where agents are opaque, heterogeneous, may be competing, and may have conflicting goals; -fault-tolerant: agents may not conform to the system specification, but the system should maintain operation, and demonstrate autonomic recovery; -volatile-tolerant: agents may join and leave the system, but the 'system' itself remains recognisably the same even if all the components change; -accountable: who performed which action, and to what effect, is significant, so social relations like trust, reputation, responsibility, liability and sanction are all significant; -decentralised: there is no central mechanism for either knowledge or control, no agent is guaranteed to have full knowledge of the entire system or control over the behaviour of all other components; -ruled by law: there is a theoretical limit on those making decisions affecting the constraints and/or requirements of behaviour of other components; -mutable: there is a mechanism by which the specification itself can be changed by the expressed consent of the participants.…”

“…elections and voting protocols) [4,5] -Rules of Social Order -Where the system characterises the permissions, obligations and (institutional) powers of each agent (eg. rights to system resources) [1,5,6] We start from a scenario with multiple agents providing/consuming resources to/from a central repository. However, the set of resources requested is more than those available for distribution, so we define a set of social constraints which determine which agent is allocated resources.…”

Competence Checking for the Global E-Service Society Using Games

“…Nevertheless, not all applications need to be presented in this way. For example, in [1] valid acts are treated in a way that relies on a more elaborate representation of concepts such as those of obligation and permission. Investigation of these aspects will allow us to compare our framework with existing approaches that model web-services, e.g.…”

“…For social interactions using agents such as an auction this differentiation will allow the auctioneer to determine which bids are valid and therefore, which bids are eligible for winning the auction [1]. In our games framework, we represent valid moves as:…”