Design and Verification of Trusted Collective Adaptive Systems
本文提出了一个形式化框架，用于建模和验证可信集体自适应系统，通过等价性和模型检测等自动化技术预测系统行为并验证信任与声誉模型，以评估系统对典型攻击的鲁棒性。
Collective adaptive systems (CAS) often adopt cooperative operating strategies to run distributed decision-making mechanisms. Sometimes, their effectiveness massively relies on the collaborative nature of individuals’ behavior. Stimulating cooperation while preventing selfish and malicious behaviors is the main objective of trust and reputation models. These models are largely used in distributed, peer-to-peer environments and, therefore, represent an ideal framework for improving the robustness, as well as security, of CAS. In this article, we propose a formal framework for modeling and verifying trusted CAS. From the modeling perspective, mobility, adaptiveness, and trust-based interaction represent the main ingredients used to define a flexible and easy-to-use paradigm. Concerning analysis, formal automated techniques based on equivalence and model checking support the prediction of the CAS behavior and the verification of the underlying trust and reputation models, with the specific aim of estimating robustness with respect to the typical attacks conducted against webs of trust.