Function Optimization Benchmarking
AI method coursework function optimization algorithm MATLAB code and datasets
Usage
Benchmarking
-
uncomment and comment different algorithms in exampleexperiment
-
generate data with 2010 BBOB framework by running exampleexperiment or dedicated experiment script in MATLAB
- set path that points to fgeneric as "pwd" e.g.
addpath(pwd)
- set fgeneric
delta precision = 1e-16
- set
datapath = '../Datasets/ALGONAME'
- set
maxfunevals = '5e3'
- set random generator seed
rng('default')
- set path that points to fgeneric as "pwd" e.g.
-
post-process data with numbbo framework
python -m cocopp [-o OUTPUT_FOLDERNAME] YOURDATAFOLDER [MORE_DATAFOLDERS]
Example command
cd Datasets python -m cocopp -o "..\Post_Processing\PSO_ppdata" "PSO"
Note
Path to datasets will be used as algorithm name in ppdata, this is why we cd into the datasets
Generate Excel Score
Results
Post Processing
Excel Scores
All tested algorithm with their corresponding scores in 5d
Algo | Score 5d | In Report |
---|---|---|
CMAES | 2344 | 0 |
DEAE | 2322 | 1 |
JADEb | 1941 | 0 |
JADE | 1868 | 1 |
JADEctpb | 1868 | 0 |
DE | 1083 | 1 |
DEb | 1083 | 0 |
PSO_Bounds | 500 | 1 |
PSO | 234 | 1 |
Bin_GA | -14 | 1 |
PSO_EDA | -49 | 1 |