def simulated_annealing(route, roads, init_temp):
current_best_distance = distance_check(route, roads)
temp = init_temp
final_temp = init_temp / 100
story = []
story.append(current_best_distance)
#if the current best distance is already optimal the loop will be skipped entirely
swap_roof = len(route) - 2
swapped_point = rnd.randint(0, swap_roof)
while temp >= final_temp * 2:
alt_route = next_distance(route, swapped_point, roads)
cost = alt_route[1] - current_best_distance
if cost < 0:
story.append(alt_route[1])
current_best_distance = alt_route[1]
route = alt_route[0]
swapped_point = rnd.randint(0, swap_roof)
temp = cool(temp, final_temp)
elif rnd.uniform(0, 1) < acceptance_probability(cost, temp):
story.append(alt_route[1])
current_best_distance = alt_route[1]
route = alt_route[0]
swapped_point = rnd.randint(0, swap_roof)
temp = cool(temp, final_temp)
else:
if (swapped_point < swap_roof):
swapped_point += 1
else:
swapped_point = rnd.randint(0, swap_roof)
temp = cool(temp, final_temp)
return story
def score_pop(Population, roads):
for gene in Population.genes:
gene[0] = distance_check(gene[1], roads)
# sorts the population
Population.genes = sorted(Population.genes, key=lambda x: x[0])
if not Population.scored:
Population.scored = True
return Population
def hill_climb(optimization_goal, route, roads):
current_best_distance = distance_check(route, roads)
story = []
story.append(distance_check(route, roads))
#if the current best distance is already optimal the loop will be skipped entirely
while current_best_distance > optimization_goal:
# keep tabs on the part of the list you're messing with
swapped_member = 2
while swapped_member < int(len(roads)) - 3:
alt_route = next_distance(route, swapped_member, roads)
if alt_route[1] < current_best_distance:
current_best_distance = alt_route[1]
story.append(alt_route[1])
route = alt_route[0]
break
else:
swapped_member += 1
if swapped_member >= int(len(roads)) - 3:
optimization_goal = maxsize
return story
def steep_climb(optimization_goal, route, roads):
current_best_distance = distance_check(route, roads)
story = []
story.append(current_best_distance)
while current_best_distance > optimization_goal:
alt_route = best_distance(route, roads)
if alt_route[1] < current_best_distance:
story.append(alt_route[1])
current_best_distance = alt_route[1]
route = alt_route[0]
else:
optimization_goal = maxsize
return story
def steep_simulated_annealing(route, roads, init_temp, temp_reduction):
current_best_distance = distance_check(route, roads)
temp = init_temp
final_temp = init_temp / temp_reduction
story = []
story.append(current_best_distance)
#if the current best distance is already optimal the loop will be skipped entirely
while temp >= final_temp * 2:
alt_route = best_distance(route, roads)
cost = alt_route[1] - current_best_distance
if cost < 0:
story.append(alt_route[1])
current_best_distance = alt_route[1]
route = alt_route[0]
elif rnd.uniform(0, 1) < acceptance_probability(cost, temp):
rnd.shuffle(alt_route[0])
explored = alt_route[0]
story.append(distance_check(explored, roads))
current_best_distance = distance_check(explored, roads)
route = explored
temp = cool(temp, final_temp)
return story