Skip to content

Latest commit

 

History

History
92 lines (67 loc) · 4.51 KB

README.md

File metadata and controls

92 lines (67 loc) · 4.51 KB
Raw

threecomphyd

This package provides tools to model performance of athletes. The name threecomphyd stands for Three Component Hydraulic and describes the performance modeling approach in focus.

Example scripts in the GitHub repository serve as instructions on how to use this package. We provide a simulation agent and example simulations, functionalities for debug plots, and an evolutionary fitting process to fit a model to individual athletes.

If you are interested in comparing this model to established performance models for intermittent exercise, you might also be interested in our pypermod project

Publication

Please see the corresponding paper on ACM or the prerpint on ArXiv for further references. If you make use of this project, we would be grateful if you star the repository and/or cite our paper.

@inproceedings{10.1145/3449726.3459469,
      author = {Weigend, Fabian Clemens and Siegler, Jason and Obst, Oliver},
      title = {A New Pathway to Approximate Energy Expenditure and Recovery of an Athlete},
      year = {2021},
      isbn = {9781450383516},
      publisher = {Association for Computing Machinery},
      address = {New York, NY, USA},
      url = {https://doi.org/10.1145/3449726.3459469},
      doi = {10.1145/3449726.3459469},
      pages = {325–326},
      numpages = {2},
      keywords = {optimization, performance modeling, metabolic response modeling},
      location = {Lille, France},
      series = {GECCO '21}
}

Setup

Please use Conda to run and develop this code. Fitting the hydraulic model requires PyGMO. It is recommended to install Pygmo with Conda and Python >3.6 and <3.10.

Usage Hints

Some hints on how to handle provided example applications in the example_scripts directory.

Tkinter on Linux Systems

On Linux systems make sure tkinter for python3 is installed via the package manager to be able to see plots and animations.

SciView in Pycharm-Professional

If you use pycharm professional, please ensure that you deactivate the SciView tab when running interactive animations.

File => Settings => Tools => Python Scientific => Uncheck "show plots in tool window"

Set Data-Storage Directory

The scripts pygmo_fittings_report.py, grid_search_evolutionary_parameters.py, and evolutionary_fitting.py read and write files to the system. They will use the location indicated in config.py to write to and read from. By default this is the location where you run the scripts from, but you can change it by setting the data_storage entry. Check in the example scripts how to do it.

Example Applications

Following demo applications are available via scripts in the example_scripts directory

  • interactive_simulation.py lets you experiment with an exemplary three component hydraulic agent and investigate its responses to various power demands.

  • model_behaviour_plots.py recreates the energy expenditure and recovery plots of the results section of the paper.

  • pygmo_fittings_report.py Iterates over all evolutionary fitting results stored in the data-storage directory defined in config and creates a readable .json overview over best fitting results, configurations, algorithm parameters, execution time, etc.

WARNING! the following scripts require a system with a minimum of 7 CPU cores for the evolutionary fitting, or even 21 CPU cores for the grid search approach.

  • grid_search_evolutionary_parameters.py starts a grid search over described parameter settings for MOEA/D coupled with the asynchronous island model by Pygmo. One fitting per parameter combination as summarised in Table 3 of the Appendix are estimated. Results are stored into a data-storage folder in the root directory of the project.

  • evolutionary_fitting.py Runs 10 fittings on an athlete using the in the script defined parameters for CP, W' and recovery rations by Caen et al. The script uses the published set of best fitting parameters (30 generations, 40 cycles, 64 population size, 7 or 14 islands) to fit the model with the outlined evolutionary computation approach (see Appendix B and C). Results are stored into a data-storage folder in the root directory of the project.