Contact  |   Impressum  

You are here: ILREconomic Modeling of Agricultural SystemsResearch ABMSIM

ABMSim - A flexible framework for Agent Based Models to simulate spatially explicit structural change in agriculture



Example from a landscape near Cologne in the ABM,
with agricultural plots, settlements and forest cover

ABMSim is a reserach activity to analyze structural change in farming in a spatial explicit setting. Drawing on earlier and on-going research activity related to structural change in agriculture and location specific analysis of cost functions in dairy farming, it aims at combining a detailed mathematical programming model for dairy farms based on a meta-modeling approach with an Agent Based Model (ABM) which covers both farmers and dairies.

It reflects an increasing research focus of our team on questions related to system dynamics, spatial explicit analysis and heterogenous agents, as seen e.g from the development of an ABM for communal livestock production systems in rural Thaba Nchu in the Free State, South Africa in the context of a DFG funded research unit or the work on the 1x1 km layer in CAPRI.

So far, a protoype version has been developed since October 2013 which comprises the basic processes to be analyzed with the Agent Based Model as well as the interface to land use cover maps, but still misses important details, e.g. with regard to the spatial competition between dairies for delivery contracts. Equally, only preliminary tests had been conducted to estimate a dual profit function from simulations with FARMDYN to provide a meta-model to be incorporated in the ABM.

The resarch activity is receiving since November 2014 funds from the German Science foundation in the context of the project Understanding spatial interactions and structural change in the dairy production chain.

Core model characteristics:

  • Set up to allow simulations with thousands of farmers and large landscapes
  • Allows to dis-aggregate the farming population in groups by specialization and with different behavior, e.g. with regard to farm exit decisions or bidding behavior in land markets
  • Considers different plot sizes and a differentiation between different types of land cover (arable and grass lands,forest, housing, other urban fabric, water bodies, other types of land
  • Landscape generated from CORINE land cover data, resolution 100x100m, landscape size >> 100 sq km
  • Comprises a simple cellular automaton based land use cover change module to account for urbanization
  • Bidding behavior of farmers for plots and milk delivery contracts can be based on a dual profit function
  • Farm structural statistics at commune level can be integrated into framework
  • Dual profit function will be estimated from simulation experiments with FARMDYN
  • Java based (1.8 using parallelism) with a Graphical User Interface implemented in GGIG, to allow for a flexible steering of the modeling system and use a well developed set of exploitation tools (maps, graphs, tables)

Documents related to the project

Methodological and technical documentation of the prototype version

ODD+D based documentation of the main characteristica of the model

Documentation of the java code (javadoc)

FARMDYN model documentation

Staff at ILR working on the project:

Publications related to the project

    Britz, W., Wieck, C. (2014): Analyzing structural change in dairy farming based on an Agent Based Model Download....

Last updated: Friday, April 07, 2017