This repository contains the configuration files, simulation routines and plotting scripts for the paper A PDE-Constrained Generalized Nash Equilibrium Approach for Modeling Gas Markets with Transport by Grimm et al.

Setup

To install the required dependencies for the julia code, from the toplevel directory run

julia --project=./julia
] instantiate

Follow the instruction at https://github.com/jump-dev/Gurobi.jl to setup Gurobi.jl

Running the simulations

In each directory under cfgs, run

julia --project=../../julia ../../julia/src/run.jl ./*

to generate the simulation results (stored as hdf5 files). Each subdirectory contains the following files:

• cfg.json contains the physical characteristics of the pipe, the market data and the operational steady state.
• solver_opts.json contains the solver options
• model_opts.json contains the penalization parameter values as well as some discretization choice.
• solver contains the name of the solver
• nb_x_pts contains the number of (discretization) nodes along the pipe.

Generating the figures

The required python packages include matplotlib, numpy, colorama.

To generate the figures displaying the various quantities of interest, run

make -f $(git rev-parse --show-toplevel)/python/figures/Makefile

from the directory containing the hdf5 files of interest.

To compare the evolution of the economical quantities with different number of players, run

python $(git rev-parse --show-toplevel)/python/figures/fig-econ-vals.py --dirs cfg-price-3-firms,cfg-price-perf-comp

from the cfgs directory.