QUANTICOL - CaSL at work

This deliverable reports on the work completed in the final reporting period on the modelling language at the centre of the QUANTICOL framework, C ARMA . A major focus of the period has been on making modelling with C ARMA accessible to a wide audience of potential users interested in CAS, not just those already familiar with formal modelling with process algebras. To this end we have further developed the C ARMA Specification Language ( CaSL ) and the software tools that support it; we have developed exemplar models, some of which are reported in this deliverable, and extended the suite of tools to offer a modeller different approaches to model analysis. In essence, the CaSL language does not formally extend the expressiveness of C ARMA , but it presents a more programmatic style of modelling, which will be usable by a wider set of people. Space plays a key role in many CAS and we have revisited the support that is offered to faithfully capture the spatial aspects within a model, resulting in improved syntax to assist the modeller and a graphical front end which can be used to automatically generate the spatial aspects of models. In this document we give an account of the key features of CaSL , and present a full account of the language in an appendix. We also discuss a few of the models that have been developed alongside the language development. These served to refine our ideas on how best to support spatial modelling as well as testing the implementation in the Eclipse plug-in supporting CaSL . One of this set of models is completely outside the domain of smart cities, which has been the main focus of our case studies, in order to demonstrate that the modelling pathway that we have developed could be suitable for a wide class of applications beyond smart cities. Of course, to be practically useful a modelling language must be implemented in a robust set of software tools to allow the modeller to construct and analyse the model with confidence. In Section 3 we give an account of the software tools, whilst in Section 4 we describe the design workflow and analysis pathway that is supported in the tools. This takes into account the different phases that a model goes through, from initial design, elaboration, parameterisation and then use as a tool to investigate the behaviour of the system under study. The modeller needs different support at each of these stages and we have sought to provide what is appropriate for each stage, as far as is feasible within the limited time and resource of the project. Building within Eclipse has allowed us to provide many "hidden" features which nevertheless greatly enhance the support for the modeller. These internal checks seek to ensure that CaSL models are free from the type of minor error that can be frustrating and time-consuming during model development. However, once a modeller is fully confident of their model, a graphical user interface can become cumbersome and inconvenient. Thus we also provide a command line interface to support efficient exploitation of models under different experimental frames. Moreover, the results of model analysis are automatically enhanced with metadata to assist with their interpretation and reproducibility. The deliverable concludes with a demonstration of the analysis of CaSL models on two of the scenarios from our smart city case studies. Specifically we consider a mesoscale model of buses within a city, particularly paying attention to the congestion that occurs when multiple routes share the same bus stopes, and issues related to regulatory compliance and appropriate spacing on frequent bus services. In the second example, we consider the key issue related to user satisfaction within urban bike sharing systems - whether a user will find a bike or a slot at a convenient location when they want one. In addition to this document, we also deliver the software tool suite for C ARMA which is available at https://quanticol.github.io