Comparison of Randomized Optimization Methods

Before You Run the Files

1. This project uses a modified version of ABAGAIL, located in the ABAGAIL sub-folder
2. The folders NNOUTPUT, CONTPEAKS, FLIPFLOP and TSP must be created in the same folder as the Jython code before running it.
3. The files sample_orig_2016.txt, sample_svcg_2016.txt must be in the same folder as the .py files
4. To run the JAVA files, you must have jpype installed. Do this before running the code in terminal / command line:

• git clone https://github.com/originell/jpype.git
• cd jpype
• python setup.py install'

Additional reference can be found here: https://stackoverflow.com/questions/35736763/practical-use-of-java-class-jar-in-python

Explanation of Files

The zip folder includes the following files:

Python code files

• NN0.py: Code for Backpropagation training of neural network
• NN1.py: Code for Randomised Hill Climbing training of neural network
• NN2.py: Code for Simulated Annleaing training of neural network
• NN3.py: Code for Genetic Algorithm training of neural network
• continuouspeaks.py: Code to use Randomised Optimisation to solve the Continuous Peaks problem
• tsp.py: Code to use Randomised Optimisation to solve the Traveling Salesman Problem
• flipflop.py: Code to use Randomised Optimisation to solve the Flip Flop Problem
• optimization_analysis.py: Code to do plotting and computation of summary statistics

Data files

• sample_orig_2016.txt
• sample_svcg_2016.txt

Resutls folders

• NNOUTPUT: Output folder for the Neural Network experiments
• CONTPEAKS: Output folder for the Continuous Peaks experiments
• FLIPFLOP: Output folder for the Flip Flop experiments
• TSP: Output folder for the Traveling Salesman Problem experiments
• ABAGAIL: folder with source, ant build file, and jar for ABAGAIL

Report

• jhung34-analysis.pdf

License

The content of this repository is licensed under a Creative Commons Attribution License