This project contains a Python class, called EvoPoolOpt, that uses an evolutionary algorithm to optimize the distribution of cities inside sport pools, using NumPy only. An instance of the class can be easily initialized with a list of city names (their GPS coordinates will automatically be downloaded), the number of pools desired and the number of chromosomes per population for the genetic algorithm.
In addition, a Jupyter notebook is provided to demonstrate a real use case, on the cities of the 2022/2023 french National 3 volley-ball division.
Create a list of cities :
cities = ['Saint avertin', 'Tarascon', 'Nantes', 'Puygouzon', 'Toulouse', 'Talence', 'Balma', 'Reze', 'Chatellerault', 'Tours',
'Montaigu Bouffere', 'Bourges', ...]Initialize an instance of the EvoPoolOptimizer with the desired cities, the number of pools and the number of chromosomes of the genetic algorithm.
opt = EvoPoolOpt(cities_list=cities, nb_pools=6, nb_chromosomes=100)The coordinates of each city will automatically be downloaded. You will only have to start start the optimization loop with the genetic algorithm by specifiying a number of iterations:
opt.optimize(num_iterations=num_iterations)Then you can observe the final result with this line :
opt.display_map_pools_configuration(opt.best_chromosome)