Self-organizing Trusted Communities (Top-down)

OC Trust调查信任和声誉等社会概念在开放式自组织系统中的使用情况。开放代理系统和开放分布式系统由于越来越多的基于互联网的应用程序的实现而受到越来越多的关注。例如桌面网格计算、无线传感器网络或自组织网络(SON)。由于这类系统的高度波动性，自组织方法是有利的。这里提交的项目涉及技术系统内社会机制的应用(系统对系统的信任)。为此，我们提出了可信社区(TC)的概念。TC由代理组成，这些代理部分自主地做出决策。他们的关系是合作的，也是竞争的。由于这些系统是开放的，我们不能假设所有的代理人都是仁慈的。TC内部的强大信任关系允许进行高效的交互，而无需重复检查和任务重复。奥委会采取了一个以分权为主的组织。这样的系统具有很好的伸缩性，并能适应当地情况。然而，他们可能会遭受不想要的紧急影响。利用积极涌现同时控制不想要的消极涌现(“受控涌现”)一直是OC研究的长期目标。这是一揽子紧急工作(EM)的目标。在三个项目阶段，我们为代理配备了越来越多的社会意识。同时，更高级别的控制程度和干扰程度也在逐渐增加。中心研究假设是：1.包含关于其他主体行为的信任相关信息提高了复杂技术系统的有效性和效率。可信代理社区在运行时的自组织可以是分散的，也可以借助更高级别的机制(制度控制)来实现。受控涌现的OC问题可以通过自主的社会主体来解决，其决策逻辑被上级机构修改。在项目第一阶段和第二阶段对纯粹分散的可信社区进行调查后，我们计划在第三阶段借助规范和制裁实现制度控制。在本项目中，我们主要关注整个系统方面(自上而下)。社会主体是在Hähner教授(自下而上)的团队中研究和开发的。我们使用的OCU中间件来自Ungerer教授的工作组。TC概念实际上是在桌面网格计算系统和智能相机网络的背景下进行评估的。此外，我们将把这一概念应用于可信能源网格(工作组Reif教授)和可信显示网格(工作组Andr6教授)。

OC TRUST investigates the usage of social concepts like trust and reputation for open selforganising systems. Open agent systems and open distributed systems are of growing interest due to the increasing internet-based implementation of many applications. Examples are Desktop Grid Computing, Wireless Sensor Networks or Self-organising Networks (SON). Due to the high volatility of such systems, self-organising approaches are advantageous.The project submitted here deals with the application of social mechanisms inside technical systems (system-to-system trust). For this purpose we have developed the concept of Trusted Communities (TC). A TC consists of agents, which take decisions partially autonomously. Their relationship is cooperative or competitive. Since these systems are open we cannot assume that all agents are benevolent. A strong trust relationship inside the TC allows for efficient interactions without double-checking and task reduplication. OC assumes a predominantly decentralised organisation. Such systems scale well and adapt to local situations. However, they can suffer from unwanted emergent effects. It has been a long-term goal of OC research to utilize positive emergence while controlling unwanted negative emergence ("Controlled Emergence"). This is the goal of the work package Emergence (EM). During the three project phases we equip the agents with increasing social awareness. At the same time the degree of higher-level control and also the degree of disturbances are gradually increased.The central research hypotheses are:1. The inclusion of trust-relevant Information about the behaviour of other agents increases the effectiveness and efficiency of complex technical systems.2. The self-organisation of a trustworthy agent community at runtime can be achieved decentralized or with the help of higher-level mechanisms (institutional control).3. The OC problem of Controlled Emergence can be solved through autonomous social agents whose decision logic is modified by higher-level institutions.After the investigation of purely decentralised Trusted Communities in the project phases 1 and 2, we plan to realize institutional control with the help of norms and sanctions during phase 3. In this project we are interested predominantly in the total system aspect (top-down). The social agents are investigated and developed in the group of Prof. Hähner (Bottom-up). We use the OCu middleware from the workgroup of Prof. Ungerer. The TC concept is practically evaluated in the context of a Desktop Grid Computing system and a Smart Camera network. In addition we will apply the concept to the Trusted Energy Grid (workgroup Prof. Reif) and the Trusted Display Grid (workgroup Prof Andr6).