GitHub

How to use the code

Setting up the repository

Note that this code was developed a few years ago and only tested on Julia 1.9.2 and with SDDP.jl 1.6.6. From today’s perspective, the Julia version and many of the packages that are used may be outdated.

We have not managed yet to update the code in order to guarantee compatibility with all newer package versions. However, when we used Pkg.instantiate() to set up our repository on a different machine, we have only experienced some issues with newer versions of SDDP.jl. For SDDP.jl it seems that at least the set_objective function has changed compared to the one we use.

Further note that running loglinearSDDP for our hydrothermal scheduling problem requires an installation of Gurobi together with a valid Gurobi license. If this requirement is not satisfied, the solver has to be changed in the run-files and a few adjustments have to be made in the code.

Running the code

Once the repository is set up, the code for our hydrothermal scheduling problem can be run using the following steps.

  • Open Julia (1.9).
  • Use cd to navigate to the repository directory.
  • Go to package mode and execute activate ..
  • It is possible that the Gurobi path has to be made known. If so, use ENV["GUROBI_HOME] = [path] where [path] is the file path of your Gurobi installation.
  • It is possible that the Gurobi package has to be built for first usage. If so, use package mode and build Gurobi.
  • Use cd to navigate to examples/Hydrothermal.
  • Execute run_model.jl (or other run-files) with include("run_model.jl").

Before using these steps, you have to make sure that the file paths in run_model.jl, run_model_linearized.jl, run_model_markov.jl, markov.jl and create_inflows_for_Markov_loglinear.jl are set as intended. The places are highlighted by TODO in the code.

This page was generated using Literate.jl.