isw4/ML02-Randomized_Optimization

CS 7641 Assignment 2 - Randomized Optimization

Code and data can be found at: https://github.com/isw4/ML02-Randomized_Optimization

ABAGAIL library modified and built from: https://github.com/pushkar/ABAGAIL
Code adapted from: https://github.com/JonathanTay/CS-7641-assignment-2

Directories

data/ Contains the raw and cleaned csv data files, along with some description of each set
graphs/ Contains graphs that are output by the various experiments
ABAGAIL/ Contains the ABAGAIL src and jar files (minor edits made from the original source)
COUNT_logs/ Contains the csv logs of the count ones problem optimizations
KNAP_logs/ Contains the csv logs of the knapsack problem optimizations
TSP_logs/ Contains the csv logs of the traveling salesman problem optimizations
NN_logs/ Contains the csv logs of the neural network optimization

Setup to run the experiments

Experiments are run on jython, which is a Java implementation of Python 2.7

Make sure to have JDK installed (jdk11: https://www.oracle.com/technetwork/java/javase/downloads/jdk11-downloads-5066655.html)
Make sure to have jython installed (https://www.jython.org/)
Add the absolute path to the ABAGAIL jar file to "ABAGAIL_absolute_path.txt". e.g.
```
C:/Documents/ML02-Randomized_Optimization/ABAGAIL/ABAGAIL.jar
```

Setup to plot the data from the logs

Data processing and plotting is run on Python 3.6, since matplotlib does not have support for Python 2

Make sure to have Conda installed
Install the conda environment:
```
conda env create -f environment.yml
```
Activate the environment: If using Windows, open the Anaconda prompt and enter:
```
activate Homework2
```
If using Mac or Linux, open the terminal and enter:
```
source activate Homework2
```

Running the datasets

Data processing:

In Conda env (Python 3):

python data_processing.py

Running Experiments:

In Jython (Python 2):
To run the neural network training experiments:

jython train_nn.py

To run the 3 optimization problems:

jython problem_countones.py
jython problem_knapsack.py
jython problem_tsp.py

Plotting

In Conda env (Python 3): Does not automatically show the plots, but will save to the graphs folder

python plot_nn.py
python plot_countones.py
python plot_knapsack.py
python plot_tsp.py

isw4 / ML02-Randomized_Optimization