Community Research and Development Information Service - CORDIS

H2020

GRACeFUL Report Summary

Project reference: 640954
Funded under: H2020-EU.1.2.2.

Periodic Reporting for period 1 - GRACeFUL (Global systems Rapid Assessment tools through Constraint FUnctional Languages)

Summary of the context and overall objectives of the project

The GRACeFUL project will lay a technological base for the building of future Rapid Assessment Tools (RATs). By this term we denote active decision-support systems for collective policy making concerning “global systems”. The introduction of RATs will significantly improve the policy-making process whenever multiple stakeholders are involved in multi-disciplinary, global challenges. A new paradigm will be designed, adopting and adapting the socially-inspired discipline of Group Model Building (GMB).
In GMB-like group sessions, a RAT will provide visual and dynamic support for the interaction between the human group process and the existing scientific evidence. The group process is the course of discussions and actions by which a collective vision is developed of the problem at hand, and at the same time a vision of the targets and actions needed to solve that problem. Problem solving and specification are intertwined in the process. Digested scientific evidence, like facts about a real-world situation and knowledge to interpret these, is used to rapidly assess the expected effects of actions and the extent to which they solve a particular problem.
Future RATs will employ solver engines to rapidly propagate and visualize the effects of stakeholders’ targets, interests and constraints. Underneath a visual and interactive shell, the bridge between the required social interaction and the underlying body of knowledge and data will be achieved through multi-level, hierarchical modelling with a novel domain-specific language (DSL) of policy concept maps, logic and relations. Embedding this DSL into a programming language combining Functional Programming (FP) and Constraint Programming (CP) would constitute a major breakthrough. On the one hand, the FP perspective empowers embedded DSL construction and improves scalability, verifiability and correctness of the models, while on the other hand the CP approach introduces goal-directed problem-solving power, which will reduce the number of needed trial-based simulation cycles.
The overall new paradigm, materialized through the use of socially interactive RATs on top of novel DSL technology, will democratize global systems. Management, understanding and modelling of complex systems will be streamlined and opened up to diverse groups of stakeholders ―experts as well as non-experts―, hence promoting collective awareness. Although we will employ a selected problem case (Climate Resilient Urban Design (CRUD) in the city of Dordrecht, NL) as a test bed, our long-term vision and approach to RAT design and building will be entirely generic, in order to maximize future innovation opportunities for numerous consultants, tool developers, and policy makers.
Specific Objectives:
This project aims at achieving the following tangible objectives:

(1) To establish a novel, human-centred, ICT-supported, collective policy-making paradigm for Global Systems Science (GSS) in which experts, stakeholders, politicians and other citizens involved in a multi-disciplinary global challenge, can participate in face-to-face problem-solving sessions. This envisaged methodological framework of Group Policy Making will lie at the junction between organizational GMB, simulation and problem solving, and will be formalized in broadly applicable scripts, guidelines and tools.
(2) To carry out foundational scientific work to bridge low-level physical and mathematical models with high-level problem narratives. This will be achieved through the construction of a DSL for system dynamics models on top of a host language that combines the CP and FP paradigms and provides multi-scale constraint and solver models. Major work will be carried out in the study of ways to compose constraints, Global Constraints , heuristics, and solvers into higher-level language constructs for the DSL.
(3) To design a graphical and interactive layer for supporting the application of (1) coupled to the DSL from (2). This will include mod

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

One of the major challenges faced with during the first reporting period was to establish a common understanding of the problem and of the technological capabilities, and to make the multidisciplinary consortium speak and understand a common language. To this end, Climate-Resilient Urban Design (CRUD) and its state-of-the-art was thoroughly analysed and made common ground for the consortium. An important GMB event was organized on CRUD involving some 80 engineering and water-management students. From the analysis, key requirements for future Rapid Assessment Tools were synthesized, mostly in the form of ontologies for CRUD and for GRACeFUL Concept Maps, name given to the basic knowledge structure that will be manipulated in RATs. The ontologies provide a common vocabulary and semantics, so that all parties give a same interpretation to the meaning of the concepts. They constitute the basic common knowledge ground in the project and will evolve and continuously made up to date.

Foundational scientific work was mostly carried out in the technological work packages WP4 and WP5. In WP4 the research has been mainly focused on applying the state-of-the-art in Domain Specific Languages to the domain of CRUD and formalizing CRUD concepts in Haskell. In WP5, a first version of the Haskell-embedded DSL for CFP was designed, implemented and delivered. Additional foundational work was done on time-series constraints and qualitative spatial and temporal constraints, which will have an important role in the CRUD application. In WP5 one paper was published in an international journal and 4 papers in international conferences. Others have been or will be submitted shortly.

In WP3 the initial Visual Analytics and Exploratory Data Analysis requirements and specifications were synthesized, based on the CRUD Analysis in WP2. Research also focused on VA of temporal dynamics of complex situations based on State Transition Graphs and on VA of multivariate time series. A specific Master’s thesis was set up for the tool-user communication aspects, including interactivity and gamification.

Towards the objective of building of generic RATs most of the work is to start in the second reporting period. Some preliminary work for the DSLs is constituted by the work done in D4.1 and D5.1, as well as the ontology work in WP2.

A regular communication has been set up with the city of Dordrecht’s StadsLab initiative. The GRACeFUL consortium has been present in many of their monthly meetings. The Dubbeldam neighbourhood has been selected as a first case study location. It has been used to focus the ontology development and to define a toy case for use during the CFP platform development. Specific stakeholders for the case study have been identified, though for the first GMB sessions a more general setup was preferred using a large number of students.

The GRACeFUL website was designed and set up (www.graceful-project.eu and associated Twitter account @GRACeFULProject) as well as the Dissemination plan template. The GRACeFUL project participated in the RRI-SSH forum, in a Policy workshop at the eGOV Conference and has recently applied for the FEATart initiative. Several scientific papers have been submitted and/or published.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

"Progress beyond the state-of-the-art can be summarized, per work package, as follows:

WP2, the CRUD case:
Deliverable D2.3 on CRUD RATs Key Requirements constitutes a step beyond the state of the art, in a conceptual/prospective way, as it defines a first generic GRACeFUL Concept Map structure applicable to other design studies, and it contains a first CRUD ontology which starts specifically from the viewpoint of CRUD actions.

WP3, Visualization:
The research on Visual Analytics of temporal dynamics of complex situations based on State Transition Graphs (STG) and VA of multivariate time series constitutes progress beyond the state-of-the-art and has been successfully applied to traffic network analysis and sports analytics, two publications on which are under review.

WP4, DSL for System Dynamics models:
A computational theory of policy advice and avoidability has been developed, which clearly extends the state-of-the-art:

• Botta, N., Jansson, P., Ionescu, C., "Contributions to a computational theory of policy advice and avoidability", submitted to the Journal of Functional Programming, Special Issue on Dependently typed Programming, 2016.

WP5, Constraint Functional Programming:
Most of the foundational work in this work package constitutes progress beyond the state-of-the-art. It is summarized in the following publications:

• Beldiceanu, N., et al. "Using finite transducers for describing and synthesizing structural time-series constraints." Constraints 21.1 (2016): 22-40.
• Beldiceanu, N., et al. "A Modelling Pearl with Sortedness Constraints", Global Conference on Artificial Intelligence: GCAI 2015, Manchester, UK: Cool Press, 2015, 27-41 p.
• Arafailova, E., Beldiceanu, N., Douence, R., Flener, P., Francisco Rodriguez, M., Pearson, J., Simonis, H., "Time-Series Constraints: Improvements and Application in CP and MIP Contexts", Proceedings Thirteenth International Conference on Integration of Artificial Intelligence (AI) and Operations Research (OR) techniques in Constraint Programming, May 2016.
• Torrini, P., Schrijvers, T. (2015). Reasoning about modular datatypes with Mendler induction. In Matthes, R. (Ed.), Mio, M. (Ed.), Proceedings Tenth International Workshop on Fixed Points in Computer Science: Vol. 191. Fixed Points in Computer Science (FICS). Berlin, 11-12 September 2015 (pp. 143-157) EPTCS. This paper explores modularity of datatypes, which is particularly relevant for setting up modular domain-specific languages, where the datatypes are the abstract syntax trees of the DSL and Mendler induction is a structured way for processing such datatypes.
• Neil Sculthorpe, Paolo Torrini, Peter Mosses. A modular structural operational semantics for delimited continuations. Proceedings of the Workshop on Continuations (WoC 2015), Electronic proceedings in theoretical computer science. This paper provides a formal underpinning for delimited continuations which are a powerful language primitive (originating in functional programming) which can be used for implementing, among others, powerful search strategies for Constraint Programming.
• Two other papers (one on mixing quantitative and qualitative reasoning for Allen relations) and one about enhancing time series constraints should be submitted by mid-April at CP.

WP6, Dissemination & Impact Promotion
Some actions have been taken related to the project's societal implications, such as participation in a cross-project RRI-SSH workshop and applying to the FeatArt initiative."

Related information

Record Number: 186630 / Last updated on: 2016-07-14