MESSAGE Robust Decision Making Framework Code

This webpage provides the source code of the MESSAGE Robust Decision Making Framework which is described in the paper Risk Hedging Strategies under Energy System and Climate Policy Uncertainties, to appear in the "Handbook of Risk Management in Energy Production and Trading", edited by Raimund Kovacevic, Georg Pflug and Maria Vespucci.

GAMS Implementation

For easy adaptability the Robust Decision Making Framework version of MESSAGE, the model has been re-implemented in the widely used GAMS (General Algebraic Modeling System) language. The GAMS code is modular and consists of several parts, that is, the main model code, two files that specificy the model structure and parametrize the deterministic core model, and two additional files that contain the stochastic input to the model. A list of these files follows:

• stochastic_model_Krey_Riahi.gms - main model code
• set_definition.gms - definition of model structure
• parameter_definition.gms - deterministic model parameters
• stochastic_parameter.gms - some stochastic model settings
• stochastic_scenario.gms - sampled stochastic parameters

The ZIP archives below contain the first four files specified above combined in a single ZIP file and, in addition, a selection of stochastic_scenario.gms files with different sample sizes N. Note that model run times vary significantly between the deterministic model (few seconds) and stochastic version with larger sample sizes (N = 20000, 20-40 minutes). For running the stochastic model version with larger sample sizes a high performance Linear and/or Quadratic Programming solver (e.g. CPLEX) is required where interior point methods have shown to perform better than simplex methods.

• stochastic_model_Krey_Riahi.zip (23 kB)
  
• stochastic_32_parameters_lognormal_lhs_N-100_S-0.zip (0.12 kB
• stochastic_32_parameters_lognormal_lhs_N-1000_S-0.zip (1.2 MB
• stochastic_32_parameters_lognormal_lhs_N-2000_S-0.zip (2.4 MB
• stochastic_32_parameters_lognormal_lhs_N-5000_S-0.zip (5.9 MB)
  
• stochastic_32_parameters_lognormal_lhs_N-10000_S-0.zip (11.8 MB
• stochastic_32_parameters_lognormal_lhs_N-20000_S-0.zip (23.7 MB)
  
  

Last edited: 19 July 2013