First of all I should say that here we are considering such kind of travelling salesman problem where cities to start and to finish the salesman journey are given.
Step 1. All distances between cities are given in a matrix called weights, so each diagonal cell equals 0 and ij-cell is a distance between cities i and j.
Step 2. For making initial population we set its size pop_size
and number of generations(gen_max)In method generate_population
we generate
initial population randomly.
Step 3. Crossover. We take two random parents(paths) and apply a
crossover method for them. crossover(self, mother, father)
The number of taken chromosomes from mother and father are
generated randomly.
Step 4. Mutation Each child after crossover exposed to mutation.
in chromosome.py you can see method mutate which just swaps two
chromosomes in a gene. So yes 100% of population is mutated.
Step 5. Fitness We measure fitness of each generation by measuring
length of proposed path. In each iteration of algorithm work we are
looking for the fittest path and put into best variable.
Step 6. Selection Method selection sort new generation and
previous generation using fitness function and gives out the best
representatives of each generation and put it together updating
population array.
In a cycle we are repeating steps 3, 4, 5 and 6 for 1000 times.