RESEARCH ACTIVITIES

Modeling and Simulation

Image

The interdisciplinary study of irrigation systems requires bringing together the expertise of various researchers and stakeholders so that they build a co-constructed and shared representation of the system. This activity therefore begins with the collective design of a conceptual model, in which the system and its components, actors and dynamics are identified, so that it can become the basis of an “operational model”, i.e. a model that can be simulated, explored and played out. This construction is similar to a brainstorming activity, but methods and tools are still lacking to translate the results obtained into computer models: how to translate the concepts (actors, resources, interactions) agreed upon by participants during the conceptual modeling step to elements that can be implemented in a computer model ? This question raises three challenges:

  1. the definition of easy-to-use, flexible tools to support the design of shared conceptual models
  2. translating a loosely defined model to a formal — and operational — one;
  3. not leaving aside the participants when the translation to the formal model is done by experts;

Goals

The goal of this activity is to provide methods and tools to support participatory modeling, by the actors chosen through assessment and by scientists from different disciplines. Its main objective consists in the design and implementation of the model that will be simulated to support the exploration of management alternatives in selected scenarios.

Methodology

The modeling approach chosen is the Agent-Based Modeling (ABM) paradigm, which represents the behavior of a system as an emerging feature of computational entities (agents) in interaction in a simulated environment. This approach has become the approach of choice to represent socio-environmental systems (such as irrigation systems), especially when they require the inputs of different disciplines to be designed. The implementation of an ABM is usually done on a dedicated software platform. GAMA is chosen for its (spatial) data management, heterogeneous models coupling and participatory simulation facilities. To overcome the three issues listed above, we chose to adopt the “simulate to design” approach promoted by C. Le Page, which blurs the distinction between conceptual and formal models, and places simulations at the centre of interactions between actors, researchers and decision-makers.

Outcomes

➔ The design, implementation and distribution of a set of extensions to the GAMA platform to support participatory modeling (graphical modeling, modular modeling through the coupling of models …)

➔ The design and implementation of a set of reusable abstractions that can be instantiated to describe
irrigation systems (works, canals,…), their managers and users, and their operation.

➔ For each case study:

  • an integrated, agent-based, spatialized model written in GAML
  • the documentation of the model and its user guide
Author