A Quantitative Approach to the Design and Analysis of Collective Adaptive Systems for Smart Cities

agenda of many European Union (EU) and other international institutions and think-tanks. As urban populations grow, innovative information and communication technology (ICT) initiatives are seen by many as one of the key factors that will allow modern cities to reach or maintain a good and sustainable quality of life for their inhabitants, allowing for the timely and equitable distribution of resources. These ICT-based systems are based on decentralised and distributed designs, comprised of many autonomous and interacting entities, known as collective adap-tive systems (CAS). CAS are required to adapt their services seamlessly to the changing needs of their users, who also form an integral part of the system. They typically consist of a large number of spatially distributed, heterogeneous entities with decentralised control and varying degrees of complex autonomous behaviour. This requires the development of novel scalable analysis techniques to investigate their dynamic behaviour and support the design and operational management of a wide range of such systems. In the QUANTICOL project [1], three principal case studies drive the development of a design and analysis framework for CAS: two smart urban transportation systems (smart bus systems and bike-sharing schemes) and smart grid applications. We present some of the QUAN-TICOL research performed at ISTI-CNR. In the first year of the project, we developed several scalable analysis techniques that exploit mean field and fluid flow techniques , in combination with logic-based model-checking, to support the investigation and prediction of dynamic resource usage. Mean field techniques were originally developed in the field of statistical physics to cope with the analysis of very large scale systems composed of interacting objects such as molecules in a gas. The possibly non-linear behaviour of such systems is conveniently modelled by a deterministic approximation, i.e., the limit for an infinite number of agents, given as the solution of a set of differential equations (in the continuous case) or difference equations (in the discrete case). Their combined application with model-checking techniques provides a way to verify properties of individual entities in the context of a large system on which they depend, but also properties of the global system or combined local and global properties. An example is the study of the potential effects of user-incentives on maintaining a satisfactory distribution of bikes and empty parking slots over time. The extension of these techniques to address spatial aspects, including spatial model-checking, is a major objective of the project [2]. A …