info2excel

A small script written for AI method coursework, to speed up the process of copying delta f target values The values were retrieve directly from .info thats generated by BBOB or numbbo framework

Previous issue

Problem

retrieving negative and zeros values

precision = 1e-8   
global_best = foptimal = 50   
ftarget = foptimal + precision   
delta fitness = fbest - ftarget

Therefore, if fbest reached the global minimum, say at 50, ftarget = 50 + 1e-8

delta_fitness = 50 - 50 - 1e-8
delta fitness = -1e-8

Solution

replace values beyond the bound with threshold values

maximum_delta = 1e+3
minimum_delta = 1e-14

Installation

Windows (cmd Batch)

clone and navigate into the repository
```
	cd [REPO]
```
create virtual environment
```
	python -m venv info2excel\venv
```
activate virtual environment (a prefix of "(venv)" will be displayed in your command prompt when succesfully activated)
```
	info2excel\venv\Scripts\activate
```

install dependencies

	pip install -r info2excel\requirements.txt

Mac (terminal Bash)

clone and navigate into the repository
```
	cd [REPO]
```
create virtual environment
```
	python -m venv info2excel/venv
```
activate virtual environment (a prefix of "(venv)" will be displayed in your command prompt when succesfully activated)
```
	source info2excel/venv/bin/activate
```

install dependencies

	pip3 install -r info2excel/requirements.txt

Generate Delta ftarget Excel Score

Execution

place datasets into Datasets folder
activate virtual environment (step 3 of installation)

generate excel

	info2excel\info2excel.py -i [DATASET] 								# or
	info2excel\info2excel.py -i [DATASET] -d [DIMENSION] -o [EXCELNAME] -u [UPPER_BOUND] -l [LOWER_BOUND]
	# OPTIONAL ARGUMENTS
	# d by default 5
	# o by default "[ALGONAME]_[DIMENSION]D"
	# u by default 1e+3
	# l by default 1e-14

Example command

	info2excel\info2excel.py -i CMAES 				#or
	info2excel\info2excel.py -i CMAES -d 5 -o CMAES_5D -u 1e+3 -l 1e-14
	# both output same content : CMAES_5D.xlsx

generated excel is in ExcelScore

Note
Always remove pre-existing dataset when rerunning the experiment

teoshibin / info2excel