Book on "Trustworthy Open Self-Organising Systems"
The results of six years of research conducted in the course of the OC-Trust project have been summarised in the book on "Trustworthy Open Self-Organising Systems". Furthermore, the book includes three invited contributions by selected experts from the domain of trust in multi-agent systems. Jeremy Pitt discusses formal models of several social processes for open distributed systems, and, in a sense, removes the restriction on the social concept of trust otherwise so prominent in this book. Cristiano Castelfranchi and Rino Falcone add various other factors to the discussion on trust in self-organising, socio-technical systems. Natasha Dwyer and Stephen Marsh conclude the book by asking the interesting and relevant question whether a digital environment empowers users to proceed on their own terms.
Open systems typically occur in a wide range of applications, from virtual organisations and vehicular networks to cloud/grid computing and reconfigurable manufacturing. All these applications encounter a similar problem: how does a system component reliably complete its own tasks, when successful task completion depends on interaction and interoperation with other, potentially unreliable and conflicting, components? One solution to this problem is trust: depending on a second party requires a willingness to expose oneself to risk, and to the extent that this 'willingness' can be quantified or qualified, it can be used to inform a binary trust decision. Therefore, a formal model of the social relationship underpinning such trust decisions is essential for conditioning bipartite interactions between components in an open system. However, there are a number of issues that follow from this. For example: what is to be done when the outcome of the trust decision is contrary to expectation? Are there positive externalities that can be derived from a successful trust decision? And how can we ensure that outcomes of collective decision-making in such circumstances are, in some sense, 'correct' and/or 'fair'? Our answers to these question have been found in the formalisation of other social relations, respectively forgiveness, social capital and justice. In this talk, we will present a survey of this development of formal models of social relations, from trust to justice via forgiveness and social capital, all of which collectively approach the fundamental issue of reliable interoperation in open systems.
Ubiquitous environments are able to support users during their daily life by appropriately adapting themselves to changing contexts. However, allowing autonomous system behaviour might lead to a decrease of transparency and controllability. This affects the interaction of system units as well as the users’ experience with such an environment. Thus the aim is to find an adequate balance between autonomy and trustworthiness to ensure desirable system behaviour. This could be done by enhancing the system’s decision making processes with the concept of trust.
Future technical systems will be increasingly characterised by openness and heterogeneity of participating elements. Within the OC-TRUST project we have developed solutions to handle anomalies, disturbances, and malicious behaviour of single agents by making use of trust and reliability metrics. Based on mutual trust and reliability ratings agents can form self-organized communities, which allow for a more effective cooperation within the community while keeping out non-trustworthy agents. The overall goal is to increase the robustness of open distributed systems with low overhead.
Self-* systems are becoming increasingly complex in their organisational structures, especially when unknown heterogeneous entities might arbitrarily enter and leave the network at any time. Therefore, new ways have to be found to develop and manage them. One way to overcome this issue is trust. Using appropriate trust mechanisms, entities in the system can have a clue about which entities to cooperate with. This is very important to improve the robustness of self-* systems, which depends on a cooperation of autonomous entities. The contributions of our work are trustworthy concepts and generic self-* algorithms with the ability to self-configure, self-optimise, and self-heal that work in a distributed manner and with no central control to ensure robustness. In this demonstration, we report some results that we experimentally evolved from our TEM middleware. We address the feasibility of applying trust to the self-configuration property. This aims on the one hand to equally distribute the load of services on nodes as in a typical load balancing scenario and on the other hand to assign services with different importance levels to nodes so that the more important services are assigned to more trustworthy nodes. We also discuss the future application of our TEM middleware.
In open multi-agent systems, we can make only little assumptions about the system's scale, the behaviour of participating agents, and its environment.
Especially with regard to mission-critical systems, the ability to deal with a large number of heterogeneous agents that are exposed to an uncertain environment becomes a major concern.
A prominent example are power management systems whose paramount goal is to balance production and consumption in the presence of uncertain weather forecasts and variable demand.