IMP_READ_THIS.txt

In order to Compile and Run the program in windows
-----------------------------------------------------------------

1.Using CMD go into the folder named MachineLearning
2.In CMD type "Compilation.bat" to compile the program
3.In CMD type "Run.bat" to run the program 

In order to modify the parameters of the Genetic Algorithm
-------------------------------------------------------------------------

1. Go in the TSPLauncher.java class and modfy in the constructor the file you want to run as benchmark and the parameters
2. If you want to modify the number of loops modify it in the game engine class TSP_GA.java

 In order to modify the parameters of the Ant Colony Optimizaton Algorithm
--------------------------------------------------------------------------------------------

1.Go in the class TSP_ACO and mofiy accoridingly the parameters found in the default constructor 
2. If the number of iterations need to be modified , modfiy them from the engine method in the class TSP_ACO.

NOTE : In the write up there is the optimal parameter values in a table, for each file found inside the Machine Learning folder. Therefore it is recommended that for optimal use the parameters should be modified 
                as guided in this README file.
NOTE : javadocs are located in the folder javadocs
NOTE : Statement of completion is found in the last page of the report
NOTE : plagiarism form is in this folder
NOTE : In the main launcher , the launcher only loads the file burma14.tsp as instance , the reason being is that different instances require different optimal parameters as highlighted in the documentation. Thus 
                in order to load more instances , change the path name from the TSPLauncher.java. It is also important to check that the file being use uses either the EUC_2D metric or the GEO metric , these were both
                included as functions in the class TSP_Graph/Graph.java . Thus if a GEO instance is called , check that the constructor in the Graph class calls the setMatrixOfWeightsGeo , otherwise the setMatrixOfWeights
                euclidean must be called . 2 metrics were computed so that it was confirmed that the algorithms are not only correct for one type of metric.
NOTE: In this implementation the GA runs first and then the ACO , the result of the GA is given prior to starting the ACO not in the end