Flow-shop Scheduling with Genetic Algorithms

Solution to the flow-shop scheduling problem using genetic algorithms.

Running

The entry point for the program is the main.py file. It executes a single instance from the Flow-shop scheduling problem at a time. The following options are available:

-i INPUT_FILE, --input-file INPUT_FILE
                    Path to the file with input instances
-p POPULATION_SIZE, --population-size POPULATION_SIZE
                    Max size for the population.
-g NUM_GENERATIONS, --num-generations NUM_GENERATIONS
                    Max number of generations (iterations) to execute.
-n N_RUNS, --n-runs N_RUNS
                    Number of times the algorithm should be executed
-o OUTPUT_FILE, --output-file OUTPUT_FILE
                    Path where the output report should be stored.

For instance, to run for the tai20_5 setup, using a population of 100 individuals for 200 generations, one can use the command:

python main.py -i data/tai20_5.txt -p 100 -g 200

A few other options should be changed inside the script.

Folder structure

Each of the following folders corresponds to a step in the genetic algorithm: initialization/, fitness/, selection/, crossover/ and mutation/. Each directory comprises different strategies for the respective steps.

Note: The file results/report.pdf contains a table summarizing the results.