Running computations

To run loglinearSDDP, we use file run_model.jl. It allows to specify several runs with different parameter configurations. Those runs are executed one after each other.

For each run, we provide a random seed (used for the sampling in the forward pass), a specific log-linear model for the uncertain data and additional log-linear and linearized models that are used for the out-of-sample simulations after SDDP has terminated.

The run_model function has the following general structure:

• Define main model parameters. For out tests, we always used 120 stage, 100 realizations per stage and 2000 scenarios in the simulations.
• Set the file path (this has to be adjusted to the user’s system when trying to reproduce our results).
• Use the previously defined parameters to set up structs of type ProblemParams and AlgoParams (see definitions in Further parameters).
• Define a simulation regime that will be used for the in-sample simulation after SDDP has terminated.
• Set up the log-linear AR process by calling function set_up_ar_process_loglinear from set_up_ar_process.jl.
• Call model_definition from hydrothermal_model.jl to construct the multistage optimization problem. This requires to pass the previously defined process as an argument.
• Call the train_loglinear function from algorithm.jl to start running SDDP.
• After running SDDP, perform several simulations. For details, see simulations.
• The function extended_simulation_analysis is used to analyze and log parts of the simulation output.

This page was generated using Literate.jl.