cmlsga-jmetalpy

Python implementation of cMLSGA built on top of the jMetalPy framework.

Build instructions

Install required packages with

pip install -r requirements.txt

Run instructions

Currently there are 2 scripts: run_experiment.py and generate_visualisations.py

run_experiment

This script takes 1 input: a parameters.json file

Run the script with

python src/run_experiment.py parameters.json

Example of parameter file:

{
    "population_size": 300,
    "max_evaluations": 10000,
    "number_of_runs": 3,
    "comment": "",

    "mlsga": ["nsgaii"],
    "algorithms": ["nsgaii", "moead", "omopso"],
    "problems": ["ZDT1", "ZDT2"]
}

mlsga determines if the experiment uses the MLSGA algorithm. Use an empty array to disable using MLSGA. Use multiple algorithms here for cMLSGA. Supported algorithms in this field are:
- nsagii, nsgaiii, moead, heia, omopso, smpso, cmpso, genetic_algorithm
algorithms: List of algorithms to use in this experiment, supported algorithms are:
- nsgaii, nsgaiii, moead, ibea, spea2, omopso, smpso, cmpso, heia
problems: List of problems to use in the experiment, supported problems are:
- ZDT1, ZDT2, ZDT3, ZDT4, ZDT6
- DTLZ1, DTLZ2, DTLZ3, DTLZ4, DTLZ5, DTLZ6, DTLZ7
- LZ09_F1, LZ09_F2, LZ09_F3, LZ09_F4, LZ09_F5, LZ09_F6, LZ09_F7, LZ09_F8, LZ09_F9
- UF1, UF2, UF3, UF4, UF5, UF6, UF7, UF8, UF9, UF10
- WFG1, WFG2, WFG3, WFG4, WFG5, WFG6, WFG7, WFG8, WFG9
- IMB1, IMB2, IMB3, IMB4, IMB5, IMB6, IMB7, IMB8, IMB9, IMB10, IMB11, IMB12, IMB13, IMB14
- MOP1, MOP2, MOP3, MOP4, MOP5, MOP6, MOP7
- DASCMOP
- FDA1, FDA2, FDA3, FDA4, FDA5
- UDF1, UDF2, UDF3, UDF4, UDF5, UDF6, UDF8, UDF9
- CDF1, CDF2, CDF3, CDF4, CDF5, CDF6, CDF7, CDF8, CDF9, CDF10, CDF11, CDF12, CDF13, CDF14, CDF15
- JY1, JY2, JY3, JY4, JY5, JY6, JY7, JY8

Detailed data for each algorithm and run will be stored in the data-300pop-10000evals-3runs- directory.

generate_visualisations

This script aggregates stats from the csv file created by run_experiment.py

Pass the data folder created into the script like so:

python src/generate_visualisations.py data-600pop-30000evals-30runs-

A csv file with the filename 300pop-10000evals-3runs.csv will be created in the directory where the script was run.

Acknowledgements

Original idea by Adam Sobey

Written by Przemyslaw Grudniewski

Updates and additional ideas by Przemyslaw Grudniewski