Fortran code for the paper : Transport Policies in Polycentric Cities

moez kilani 6649f9c840 Mise à jour de 'README.md' 1 jaar geleden
bobyqa 2061e30f8b premier commit 5 jaren geleden
cobyla 2061e30f8b premier commit 5 jaren geleden
src b4685dec2f elaticity minor update 3 jaren geleden
README.md 6649f9c840 Mise à jour de 'README.md' 1 jaar geleden
makefile 7690f8b9f6 Adding makefile 3 jaren geleden
scbd.dat 93b32d4c53 second update elasticities 3 jaren geleden

README.md

Overview

This package is provided as a companion to the paper

Transport policies in polycentric cities, by Q. David and M. Kilani, Transportation Research Part A: Policy and Practice, 2022.

The package is composed of two executables :

  • scbd : fixed frequencies
  • f_scbd : endogenous frequencies

The optimization step relies on the packages BOBYQA and COBYLA written by Pr. Powell (the double precision versions are used).

The file scbd.dat is read for model parameters.

The source code is located in the src subdirectory. The subroutines are separated into several files. The EQUILIBRIUM subroutine is central since it is the one where the network equilibrium is computed.

Installation

  1. Install gfortran (or another Fortran compiler)

  2. Clone this directory

    git clone https://gogs.univ-littoral.fr/mkilani/scbd
    
  3. Check makefile (edit compiler name if not gfortran)

  4. Type make; scbd and f_scbd executables are produced

  5. Edit parameter values in scbd.dat file and save

  6. The command ./scbd can then be used and the output is written to the screen. Standard redirection can be used to save output to file.

Remark

For some parameter values the obtained output may not be optimal because the trust region search is not convenient. In case of doubt, it is preferable to change the values of parameters RHOBEG and RHOEND and check for robustness. Too small values may miss the exploration of the optimal region and values that are too large are likely to yield instability and non convergence. The given values should work well if the values of the model parameters are are not changed so much.