- Instances
** Final
- MediumSet
+ folder that saves the small-medium set of instances (
10 and that take 3060 minutes to solve by ILP) - LargeSet + folder that saves all the instances ever created , comprehends instances that take more than 60min to solve, and are around 100- Pending
- used to store instances that need to be tested somehow separated from the Medium and Large set
- Pending2, Paused,
- similar than Pending
- Test
- instances used to test the validity of the models
- Pending
- Results
- Final
- LargeSet
- MediumSet
- ILPEvolution
- GRASPvsBRKGA
- Pending
- Test
- Final
To solve instances of the problem with the models run, from ./Tools/Launcher:
For ILP (calls oplrun from commandline)
python launcher.py --solver ILP
For GRASP
python launcher.py --solver grasp
For BRKGA
python launcher.py --solver brkga
These commands solve all instances from the default folder ./Instances/Pending and save the results in a json format to ./Results/Pending Options:
Instances folder and results folder (paths must be relative to ./Tools/Launcher)
:
python launcher.py --solver grasp --instances ../../Instances/Final/MediumSet --results ../../Results/Final/MediumSet
GRASP parameters:
python launcher.py --solver grasp --iterations 5 --alpha 0.15 --ls best
BRKGA parameters:
python launcher.py --solver brkga --generations 1 --eliteprop 0.3 --mutantprop 0.2 --population 2 --inheritance 0.1 --decoder <hini,horder,hexcess>
` cd Tools/Launcher python megalauncher_grasp.py ../../Instances/Final/LargeSet_20180106/
`
Quick demo of the files and dirs created in ../../Results/Final/LargeSet_20180103/. It does not execute any solver, just time.sleep(0.5)
cd Tools/Launcher python megalauncher.py --demo ../../Instances/Final/LargeSet
Real execution
cd Tools/Launcher python megalauncher.py ../../Instances/Final/LargeSet
See results in ../../Results/Final/LargeSet_20180103/
See progress files in ../../Results/Progress/
Other options that can be used are to use the ./Metaheuristics/main.py file from it's 'main' section, calling any of the "entry points" of the Metaheuristics algorithms.
mediumSetGraph.py, takes all json files in a folder and creates 2 graphs, one comparing the solving times for differents instances for the 3 algorithms (ILP, GRASP and BRKGA). It expects to find json files with results from executions of the 3 algorithms in the indicated folder:
python mediumSetGraph.py ../../Results/MediumSet
summary_of_executions.py, example file that reads json results files from a folder and plots all the resutls in a graph
classifider.py script used to organize the executions of instances by the ILP, it saves the instances files (.dat) to the Instances/Final/ with a name that reflects the time it takes for the ILP to solve it.