import numpy
import matplotlib.pyplot as plt

def fitness_fun(indiv_chrom):
		in_sample  = [8, 4, 2, 1]
		quality = ((15*numpy.sum(indiv_chrom*in_sample)) - (numpy.sum(indiv_chrom*in_sample)*numpy.sum(indiv_chrom*in_sample)))
		return quality
	
def pop_fitness(pop):
		qualities = numpy.zeros(pop.shape[0])
		for indv_num in range(pop.shape[0]):
			qualities[indv_num] = fitness_fun(pop[indv_num, :])
		return qualities
			
			
def select_mating_pool(pop, qualities, num_parents):
		parents = numpy.empty((num_parents, pop.shape[1]))
		for parent_num in range(num_parents):
			max_qual_idx = numpy.where(qualities == numpy.max(qualities))
			max_qual_idx = max_qual_idx[0][0]
			parents[parent_num, :] = pop[max_qual_idx, :]
			qualities[max_qual_idx] = -10000
		return parents
	
def crossover(parents, n_individuals=6):
		new_population = numpy.empty((6, 4))
		
		#Previous parents (best elements).
		#new_population[0:parents.shape[0], :] = parents
		new_population[0, 0:2] = parents[0, 0:2]
		new_population[0, 2:] = parents[2, 2:]
		
		new_population[2, 0:2] = parents[2, 0:2]
		new_population[2, 2:] = parents[0, 2:]
		
		new_population[1, 0:2] = parents[1, 0:2]
		new_population[1, 2:] = parents[3, 2:]
		
		new_population[3, 0:2] = parents[3, 0:2]
		new_population[3, 2:] = parents[1, 2:]
		
		new_population[4, :] = parents[2, :]
		new_population[5, :] = parents[3, :]
		return new_population
		
def mutation(population):
		#for idx in range(population.shape[0]):
		population[0, 1] = numpy.abs(1-population[3, 1])
		return population


#population size
sol_per_pop = 6

#Mating pool size
num_parents_mating = 4

#Creating an initial population randomly
new_population = numpy.zeros((6,4))
new_population[0, :] = [1, 1, 0, 0]
new_population[1, :] = [0, 1, 0, 0]
new_population[2, :] = [0, 0, 0, 1]
new_population[3, :] = [1, 1, 1, 0]
new_population[4, :] = [0, 1, 1, 1]
new_population[5, :] = [1, 0, 0, 1]


for iteration in range(1):
	#Measing the fitness of each chromosome in the population
	qualities = pop_fitness(new_population)

	#Selecting the best parents in the population for mating
	parents = select_mating_pool(new_population, qualities,num_parents_mating)

	#Generating next generation using crossover
	new_population = crossover(parents, n_individuals=sol_per_pop)

	#new_population = mutation(new_population)
	
	qualities = pop_fitness(new_population)

in_sample = [8, 4, 2, 1]
new_population2 = numpy.zeros((6))
for k in range (6):
	print ('x = ')
	print(numpy.sum(new_population[k, :] * in_sample))
	new_population2[k]= numpy.sum(new_population[k, :] * in_sample)

print ('fitness = ')
print (qualities)

x=numpy.linspace(0,15,100)



y=15*x - x*x

plt.grid(True)
plt.xlabel('X')
plt.ylabel('Fitness F(x)')

plt.plot(x,y,"r--",label="Fitness")
for k in range (6):
    plt.plot(new_population2[k],15*new_population2[k]-new_population2[k]*new_population2[k], marker ="o",color="black")

plt.legend()
plt.show()