View-based Development of a Simulation Framework for - PowerPoint PPT Presentation

View-based Development of a Simulation Framework for Multi-Disciplinary Environmental Modelling Rolf Hennicker, Matthias Ludwig Ludwig-Maximilians-Universität München

The GLOWA-Danube Project (2000-2010) Integrative Techniques, Scenarios and Strategies for the Future of Water in the Upper Danube Basin Social Sciences Upper Danube Basin: Natural Sciences • Area: 77.000 km² • • Hydrology Environmental Psychology • Population: 8.2 Mio. • • Plant Ecology Environmental Economy • Elevation Gradient: 3300 m • • Glaciology Tourism Research • Meteorology • Water Supply • Groundwater • Agricultural Economics • Surface Water + Informatics 2

An impression from the Danube Catchement

Mutually Dependent Processes “Stand - alone” modelling of the single processes is not sufficient • An integrative view is needed → system of coupled simulation models • 4

Goal • Platform for integrative simulations with coupled models from various disciplines • Approach: Generic Simulation Framework 5

Generic Framework for Coupled Simulations Framework Ideas • Extract common properties and rules which hold for all simulation models and implement them in a general, abstract template . • The model developer must only implement the open pieces of the template (according to his/her domain). Framework Core Framework Architecture (runtime environment for configuration and coordination) Integrative Simulation System Developer Interface (e.g. DANUBIA) (“plug points” for model developers) Coupled Simulation Models

Development Principles • Different abstraction levels and refinement • Modeling with the Unified Modeling Language (UML) • Formal Methods – Object Constraint Language (OCL) – Process algebra FSP (Finite State Processes) [Magee, Kramer] • Separation of concerns 7

Politische Rahmenbedingungen [a] In-/Exfiltrationsraten Akzeptanz der politischen Stickstoff im GW [mon] Bedingungen [a] Grundwasser Politik Oberflächengew. Grundwasserneubildung mit Stickstoffgehalt [d] Wasserstand, Abfluss, N Politische Rahmen- im Flusswasser [d] bedingungen [a] Wasserstand, Abfluss, N Brauchwassermenge, im Flusswasser [h] Nutzungsart [d] Akzeptanz der politischen Grundwasserentnahme Grundwasserstand, Laterale Zuflüsse mit Bedingungen [a] Wassermenge, und Qualität [d] Stickstoff im GW [d] Stickstoffgehalt [h] Wasserqualität [h] Trinkwasserbedarf und Schnee- benötigte Qualität [d] bedeckung [d] Schneebedeckung, Abfluss aus Aktuelle Schnee- und Eisflächen [h] Landnutzung Hydrologie/ Glaziologie Psychologie Tourismus Fernerkundung Tourismusart [mon] Prognost. Landnutzung und Bewirtschaftung [a] Verdunstung, Albedo, Subjektive Risikoeinschätzung, Landnutzung [h] Akzeptanz der Wasserqualität [a] LAI, Vegetationshöhe, Wurzeltiefe, N min -Boden [d] Verdunstung, Bodenfeuchte, Albedo, Oberflächentemperatur [h] Landnutzung, Transpiration, Angebotsfunk- Nachfrage, Zahlungs- Wasserpreise [a] Albedo, Bodenfeuchte [h] tion, Preise [a] bereitschaft [a] Lufttemperatur, Wind, Feuchte, Niederschlag, Strahlung [h] Nachfrage, Zahlungs- bereitschaft [mon] Düngemittel Kosten der Wasser- Import/Export [a] aufbereitung [a] Meteorologie Pflanzenökol. Ökonomie Pol. Programme Produktion in Land- und Ökologie und Forstwirtschaft [a] Daten zu Raumordnungs- regionen [a] 8

Functional Views of the Framework • Data exchange (at runtime) • Simulation space • Simulation time and coordination 9

Outline • Introduction • View-based Framework Development • Framework Instantiation • Conclusion 10

View-based Framework Development 11

Base Requirements Design 12

Simulation Space Requirement (1) A simulation space consists of a set of ” proxels ” (process pixels). Example: Requirement (2) All simulation models, participating in an integrative simulation, agree on the simulation space. Invariant context Simulation inv: self.models −>forAll (m | m.proxels.pid −>asSet () = self.area.proxels.pid −>asSet() 13

Design Model (Space) 14

View-based Framework Development 15

Life Cycle of Simulation Models -begin -timeStep -end 16

The Coordination Problem Each simulation model • has an individual time step, • must be supplied with valid data according to its local model time Specification with FSP [Magee, Kramer] const simStart = 0 const simEnd = 6 range Time = simStart..simEnd property VALIDDATA(User, StepUser, Prov, StepProv) = VD[simStart][simStart], VD[nextGet:Time][nextProv:Time] = // no obsolete data (when ( nextGet<nextProv ) [User].get[nextGet] -> VD[nextGet+StepUser][nextProv] // no overwritten data |when not(nextGet<nextProv ) [Prov].prov[nextProv] -> VD[nextGet][nextProv+StepProv]). 17

Labelled Transition System VALIDDATA(User=1, StepUser=2, Prov=2, StepProv=3)

FSP-Design Model Model 1 ... Model n Timecontroller 19

From UML to FSP and back UML-Requirements Model FSP-Requirements Spec Formal Correctness Proof UML-Design Model FSP-Design Model 20

UML-Design Model 21

View Integration 22

Component-based Architecture 23

Outline • Introduction • View-based Framework Development • Framework Instantiation • Conclusion 24

Framework Instantiation Framework Core Developer Interface Groundwater Model 25

GLOWA-Danube Integrative Simulation System DANUBIA (Various configurations of up to 18 models) 26

Climate and Society Scenarios Integrative Simulation Evaluation of results, Discussion with Stakeholders 27

Results for the Upper Danube Basin: 2011 - 2060  Used Climate Scenario (IPCC): temperature increase 3.3°C – 5.2°C between 1990 and 2090  Trends for precipitation: More rainfall in winter, less in summer, moderate decrease all over the year  Consequences:  Expected reduction of water power production (between 10 % – 16%)  Possible restrictions for ship traffic in summer due to low water levels  30 – 60 days less snow cover per year in lower alpine regions (due to temperature increase) but possible improvements in high-level alpine regions  Less winter tourism but moderate increase of summer tourism  Further results  Less private water use expected (around 20%) due to changing behaviours and new technologies (for saving water)  Shortage of drinking water not expected, but the need for temporary adaptation strategies of water suppliers is likely (e.g. more cooperation and networks)  (Almost) all glaciers in the Upper Danube catchment will vanish until 2045 28

Characteristics of the Framework • Data exchange at runtime • Parallel execution of dependent models • Other approaches – OpenMI: Sequential execution of dependent models – Object Modeling System (OMS): Parallel execution of independent models – ModCom: no parallel execution 29

Conclusion  View-based development helped to overcome problems with system complexity; should be applicable to other types of systems as well!  The framework is generic and can be applied to any kind of (environmental) simulation model supporting: • loose coupling with data exchange via interfaces, • simulation spaces organised by proxels (of arbitrary size), • discrete time steps (of arbitrary length).  Role of Informatics in multi-disciplinary projects: Well-known methods of Informatics like abstraction, structuring , and separation of concerns can be very useful for conceptual integration .  Open DANUBIA  www.glowa-danube.de 30