Skip to main content

Global systems Rapid Assessment tools through Constraint FUnctional Languages

Periodic Reporting for period 3 - GRACeFUL (Global systems Rapid Assessment tools through Constraint FUnctional Languages)

Reporting period: 2017-02-01 to 2018-01-31

GRACeFUL lays a base for Rapid Assessment Tools (RAT), active DSS for collective policy making in global systems. RATs improve the policy-making process where multiple stakeholders are involved in multi-disciplinary, global challenges. A paradigm has been designed, adapting the discipline of Group Model Building (GMB).
In GMB sessions RATs provide support for interaction between the human process and existing scientific evidence. Digested evidence, like facts and knowledge about the real-world, is used to rapidly assess the effects of actions and the extent to which they solve a problem.
RATs employ solvers to propagate the effects of stakeholders’ targets, interests and constraints. The bridge between the social interaction and the underlying body of knowledge is achieved through modelling with a novel domain-specific language (DSL) of policy concept maps. Embedding this DSL into a programming language combining Functional Programming and Constraint Programming is a major breakthrough. The FP perspective empowers embedded DSL construction and improves scalability, verifiability and correctness of the models, while the CP approach introduces goal-directed problem-solving.
Management, understanding and modelling of complex systems is streamlined and opened up to diverse groups of stakeholders, hence promoting collective awareness. Although we focus on a selected problem, Climate Resilient Urban Design (CRUD), the approach is entirely generic, in view of future innovation opportunities.
The project aims at the following objectives:

(1) To establish a novel collective policy-making paradigm in which experts and stakeholders can participate in face-to-face problem-solving sessions. The framework will lie at the junction between GMB, simulation and problem solving.
(2) To carry out scientific work to bridge physical and mathematical models with high-level narratives. This will be achieved through the construction of a DSL for system dynamics models on top of a language that combines CP and FP and provides multi-scale solver models. Major work will be carried out in the study of ways to compose constraints and heuristics into higher-level language constructs.
(3) To design a visual layer for supporting the application of (1) coupled to the DSL from (2). This will include modes for visualizing spatial/temporal data, and manipulating cascading concept charts, causal loop diagrams, flow diagrams and constraint models. Interactivity will employ gamification to enhance user engagement.
(4) To enable the building of RATs supporting (1), using (2) and (3). This will be materialized in the form of DSL programming guidelines and a library of language constructs representing concepts common to many global challenges.
(5) To apply (4) in practice in the construction of a prototype for solving a specific real-world case of CRUD.
(6) To assess usability and user experiences of (5) and dissemination of results.
A major challenge was to get common problem understanding and to make the consortium speak a common language. For this, CRUD and its state-of-the-art were analysed and made common ground. Big student GMB events were organized. Key requirements for RATs were synthesized, mostly in the form of ontologies for CRUD and GRACeFUL Concept Maps. The common the knowledge ground allowed a better focus for the task dealing with the stakeholder GMB paradigm. Practical sessions provided useful insights to planning support tools and their use and evaluation in participatory CRUD processes. Research on value elicitation methods in the context of applying GMB to CRUD extends the s.o.t.a. in the field.
WP4 focused on applying the s.o.t.a. in DSLs to CRUD and formalizing its concepts in Haskell. A DSL for GCMs was implemented, including its formal semantics and a middleware to the underlying layer. In WP5 a Haskell-embedded DSL for Constraint Functional Programming was implemented, including a MiniZinc backend. Additional work was done on time-series constraints and qualitative spatial and temporal constraints. The work on constraints composition focused on the combinatorial aspects of time-series constraints and was made available in MiniZinc. Checkers for the latter constraints were synthesized. Focused on CRUD a constraint model for CLDs from GMB sessions was developed based on QPNs. Work was done on time and space constraints when combining actions subject to stakeholder criteria.
In WP3 the focus of effort was on visual model building for concept maps and CLDs. A prototype RAT was implemented.From the DSL side, the interface of the generic library of concepts to the visual layer was defined and implemented. The DSL for concept maps set a starting point for the construction of a growing library of generic building blocks representing concepts and relations for modelling CRUD. For the purpose of software development a toy set of blocks was defined. The library is dynamic and will be further populated with easily-definable elements.
Regular communication was set up with several CRUD case studies and practitioners (Dordrecht, Berlin, Rotterdam, New Orleans, Guayaquil). Specific stakeholders were identified, though for the GMB sessions a more general setup was preferred using students. A first toy model was developed to streamline software development. The software was validated through an expert focus group.
The website was set up (www.graceful-project.eu and Twitter @GRACeFULProject) as well as the Dissemination plan. 3 issues of the project Newsletter was published. Cross-dissemination talks were held with related projects. A cross-project joint dissemination event was organized in Brussels. The in-house dedicated dissemination tracker was used to compile inputs of the partners.
The main exploitable project results are 1) a DSL-based RAT supporting Policy decision making; 2) a decision-making process supported by RAT; 3) a Group Model Building Tool.
The CRUD RAT Key Requirements constituted a step beyond the s.o.t.a in a conceptual/prospective way, as they defined a generic GRACeFUL Concept Map structure applicable to other studies, and contained a CRUD ontology which starts specifically from actions. The research on value elicitation methods in combination with applying GMB methodology to CRUD is also considered progress beyond the s.o.t.a. as assesed by the expert focus group at the validation workshop.
All of the work published in the scientific WP4 and WP5 are to be considered progress beyond the s.o.t.a. In total there are about 30 scientific publications.
Work done in WP6 is specifically about impact promotion. The information on impact in the DoA is still valid.
GRACeFUL Overview Poster
GRACeFUL logo (vertical)
GRACeFUL logo (horizontal)