def run_with_pygame():
'''
FOR DEBUGGING PURPOSE ONLY.
First build a grid and then show every step with pygame.
This will run slowely.
For debugging purposes.
'''
import display_debug
# place houses
not_placed = True
# place 60 houses
houses_to_place = 3
while(not_placed):
try:
matrix, houselist, water = init.initialize_matrix(
houses_to_place, WIDTH, HEIGTH)
not_placed = False
except:
not_placed = True
# calculate initial profit
fs.calculate_new_vrijstand_to_class(matrix, houselist)
result = prof.calculate_value(houselist)
# save findings
state = ss.saved_state(result, houselist, water)
display_debug.build_grid(state, matrix)
def run_with_pygame():
'''
FOR DEBUGGING PURPOSE ONLY.
First build a grid and then show every step with pygame.
This will run slowely.
For debugging purposes.
'''
import display_debug
# place houses
not_placed = True
# place 60 houses
houses_to_place = 3
while (not_placed):
try:
matrix, houselist, water = init.initialize_matrix(
houses_to_place, WIDTH, HEIGTH)
not_placed = False
except:
not_placed = True
# calculate initial profit
fs.calculate_new_vrijstand_to_class(matrix, houselist)
result = prof.calculate_value(houselist)
# save findings
state = ss.saved_state(result, houselist, water)
display_debug.build_grid(state, matrix)
def showGaussian(houses_to_place, scope):
"""
Random Sampling, it will return 2 list, first the list with all values
and second a list with all free space.
ShowGaussian takes 2 parameters.
houses_to_place: the amount of houses to place 1 for 20, 2 for 40 and
3 for 60.
scope: the amount of random samples you want to take.
"""
# save found values
value_list = []
free_space_list = []
# do random sample scope amount of times
for i in range(scope):
# if a grid could not be placed it will give an error. So surround
# with try execpt to catch the error.
placed = True
while(placed):
try:
# place houses
matrix, houselist, water = init.initialize_matrix(
houses_to_place, WIDTH, HEIGTH)
placed = False
except:
placed = True
# update free space
fs.calculate_new_vrijstand_to_class(matrix, houselist)
# calculate profit
free_space = prof.calculate_vrijstand(houselist)
value = prof.calculate(houselist)
# update
if(i % 10 == 0):
print((i*100)/scope, "%")
value_list.append(value)
free_space_list.append(free_space)
return value_list, free_space_list
def showGaussian(houses_to_place, scope):
"""
Random Sampling, it will return 2 list, first the list with all values
and second a list with all free space.
ShowGaussian takes 2 parameters.
houses_to_place: the amount of houses to place 1 for 20, 2 for 40 and
3 for 60.
scope: the amount of random samples you want to take.
"""
# save found values
value_list = []
free_space_list = []
# do random sample scope amount of times
for i in range(scope):
# if a grid could not be placed it will give an error. So surround
# with try execpt to catch the error.
placed = True
while (placed):
try:
# place houses
matrix, houselist, water = init.initialize_matrix(
houses_to_place, WIDTH, HEIGTH)
placed = False
except:
placed = True
# update free space
fs.calculate_new_vrijstand_to_class(matrix, houselist)
# calculate profit
free_space = prof.calculate_vrijstand(houselist)
value = prof.calculate(houselist)
# update
if (i % 10 == 0):
print((i * 100) / scope, "%")
value_list.append(value)
free_space_list.append(free_space)
return value_list, free_space_list
def hill_climber_value(steps, houses_to_place):
'''
Finds max profit using Hill Climber. Returns a dictionary with saved data.
Takes 2 arguments.
Steps: the amount of steps a hill climber has to make. A steps is defined
as a succesful move of a house or water.
Houses_to_place: 1 for 20 houses, 2 for 40, 3 for 60.
'''
print("Starting hill climber max value")
# save profit and time
found_profit_per_run = []
time_needed = []
# 90% change to move house, 10% change to move water
water_or_house = 0.9
# place houses
not_placed = True
# start placing
while(not_placed):
print("Initialize")
init.initialize_matrix(houses_to_place, WIDTH, HEIGTH)
try:
matrix, houselist, water = init.initialize_matrix(
houses_to_place, WIDTH, HEIGTH)
not_placed = False
except:
not_placed = True
# calculate initial profit
fs.calculate_new_vrijstand_to_class(matrix, houselist)
result = prof.calculate_value(houselist)
found_profit_per_run.append(result)
# save findings
state = ss.saved_state(result, houselist, water)
# do the hillclimber
start_run = time.time()
runtimes = 0
while (runtimes < steps):
# update
if( runtimes % (steps / 100) == 0):
print(runtimes)
# make a copy of houselist and water
houselist = state.get_houselist_copy()
water = state.get_water_copy()
# random move water or house
if(water_or_house > random.random()):
house = random.randint(0, len(houselist) - 1)
moved, new_matrix = move.move_house(
matrix, houselist[house], 100, 100)
else:
water_pool = random.randint(0, len(water.get_pools()) - 1)
moved, new_matrix = move.move_water(
matrix, water.get_pool(water_pool), 10, 10)
# if a house or water has succesfully been moved.
if(moved):
# calculate new profit
fs.calculate_new_vrijstand_to_class(new_matrix, houselist)
new_profit = prof.calculate_value(houselist)
# if profit is more, accept it
if(new_profit >= state.get_total_value()):
state.set_houselist(houselist)
state.set_total_value(new_profit)
state.set_water(water)
matrix = new_matrix
# save found data
found_profit_per_run.append(state.get_total_value())
time_needed.append(time.time() - start_run)
# One step is completed
runtimes += 1
# save final profit
final_profit = state.get_total_value()
# done
end_run = time.time()
# make dictionary to return data
return_values = {"matrix": matrix, "profit_per_run":found_profit_per_run,
"time_needed":time_needed, "start_time":start_run,
"end_time":end_run, "final_profit":final_profit}
return return_values
def simulated_annealing_vrijstand(houses_to_place):
'''
Finds max vrijstand using Simulated annealing. Returns a dictionary with saved data.
Takes 2 arguments.
Houses_to_place: 1 for 20 houses, 2 for 40, 3 for 60.
'''
print("simulated annealing")
# save profit and time
found_profit_per_run = []
time_needed = []
# 90% change to move house, 10% change to move water
water_or_house = 0.9
# place houses
not_placed = True
# start placing
while(not_placed):
try:
matrix, houselist, water = init.initialize_matrix(
houses_to_place, WIDTH, HEIGTH)
not_placed = False
except:
not_placed = True
# calculate initial profit
fs.calculate_new_vrijstand_to_class(matrix, houselist)
result = prof.calculate_free_space(houselist)
found_profit_per_run.append(result)
# save findings
state = ss.saved_state(result, houselist, water)
# Start Temp Simulated Annealing
temperature = 50
start_temp = temperature
count = 0.0
# do the simulated annealing
start_run = time.time()
while(temperature > 0.08):
# update
if(count % 100 == 0):
print(count, temperature)
# make a copy of houselist
houselist = state.get_houselist_copy()
water = state.get_water_copy()
# random move water or house
if(water_or_house > random.random()):
house = random.randint(0, len(houselist) - 1)
moved, new_matrix = move.move_house(matrix, houselist[house], 100, 100)
else:
water_pool = random.randint(0, len(water.get_pools()) - 1)
moved, new_matrix = move.move_water(matrix, water.get_pool(water_pool), 10, 10)
# if a house or water has succesfully been moved.
if(moved):
# calculate new profit
fs.calculate_new_vrijstand_to_class(new_matrix, houselist)
new_profit = prof.calculate_free_space(houselist)
# if more accepts
if(new_profit >= state.get_total_value()):
state.set_houselist(houselist)
state.set_total_value(new_profit)
state.set_water(water)
matrix = new_matrix
# else random accept depending on temperature
elif((state.get_total_value() - new_profit) / temperature < random.random()):
state.set_houselist(houselist)
state.set_total_value(new_profit)
state.set_water(water)
matrix = new_matrix
# decrease temperature
count += 1
temperature = start_temp * ((0.999) ** abs(count))
# save findings
found_profit_per_run.append(state.get_total_value())
time_needed.append(time.time() - start_run)
# final profit
final_profit = state.get_total_value()
# done
end_run = time.time()
return_values = {"matrix": matrix, "profit_per_run":found_profit_per_run,
"time_needed":time_needed, "start_time":start_run,
"end_time":end_run, "final_profit":final_profit}
return return_values
def build_grid(state, matrix):
colours = get_colours()
screen = init_pygame(640, 600)
tilesize = 2 # width and height
running = True
# Start Temp Simulated Annealing
temperature = 1.5*10**6
start_temp = temperature
count = 0.0
while running:
for event in pygame.event.get():
if event.type == pygame.QUIT:
running = False
# copy of houselist
houselist = state.get_houselist_copy()
water = state.get_water_copy()
# Move house
water_or_house = 0.9
if(water_or_house > random.random()):
print("move house")
house = random.randint(0, len(houselist) - 1)
moved, new_matrix = mh.move_house(matrix, houselist[house], 10, 10)
else:
print("move water")
water_pool = random.randint(0, len(water.get_pools()) - 1)
moved, new_matrix = mh.move_water(matrix, water.get_pool(water_pool), 20, 20)
# if a house or water has succesfully been moved.
if(moved):
# calculate new profit
fs.calculate_new_vrijstand_to_class(new_matrix, houselist)
new_profit = prof.calculate_value(houselist)
# if more accepts
if(new_profit >= state.get_total_value()):
state.set_houselist(houselist)
state.set_total_value(new_profit)
state.set_water(water)
matrix = new_matrix
# else random accept depending on temperature
elif((state.get_total_value() - new_profit) / temperature < random.random()):
print("Declined yet accepted")
state.set_houselist(houselist)
state.set_total_value(new_profit)
state.set_water(water)
matrix = new_matrix
# decrease temperature
count += 1
temperature = start_temp * ((0.999) ** abs(count))
print("temp: " + str(temperature))
# draw
for row in range(300):
for column in range(320):
pygame.draw.rect(screen, colours[new_matrix[row][column]],
(column * tilesize, row * tilesize, 10, 10))
caption = "Amstelhaege. Profit: " + str(state.get_total_value())
pygame.display.set_caption(caption)
pygame.display.flip()
pygame.display.quit()
pygame.quit()
def hill_climber_value(steps, houses_to_place):
'''
Finds max profit using Hill Climber. Returns a dictionary with saved data.
Takes 2 arguments.
Steps: the amount of steps a hill climber has to make. A steps is defined
as a succesful move of a house or water.
Houses_to_place: 1 for 20 houses, 2 for 40, 3 for 60.
'''
print("Starting hill climber max value")
# save profit and time
found_profit_per_run = []
time_needed = []
# 90% change to move house, 10% change to move water
water_or_house = 0.9
# place houses
not_placed = True
# start placing
while (not_placed):
print("Initialize")
init.initialize_matrix(houses_to_place, WIDTH, HEIGTH)
try:
matrix, houselist, water = init.initialize_matrix(
houses_to_place, WIDTH, HEIGTH)
not_placed = False
except:
not_placed = True
# calculate initial profit
fs.calculate_new_vrijstand_to_class(matrix, houselist)
result = prof.calculate_value(houselist)
found_profit_per_run.append(result)
# save findings
state = ss.saved_state(result, houselist, water)
# do the hillclimber
start_run = time.time()
runtimes = 0
while (runtimes < steps):
# update
if (runtimes % (steps / 100) == 0):
print(runtimes)
# make a copy of houselist and water
houselist = state.get_houselist_copy()
water = state.get_water_copy()
# random move water or house
if (water_or_house > random.random()):
house = random.randint(0, len(houselist) - 1)
moved, new_matrix = move.move_house(matrix, houselist[house], 100,
100)
else:
water_pool = random.randint(0, len(water.get_pools()) - 1)
moved, new_matrix = move.move_water(matrix,
water.get_pool(water_pool), 10,
10)
# if a house or water has succesfully been moved.
if (moved):
# calculate new profit
fs.calculate_new_vrijstand_to_class(new_matrix, houselist)
new_profit = prof.calculate_value(houselist)
# if profit is more, accept it
if (new_profit >= state.get_total_value()):
state.set_houselist(houselist)
state.set_total_value(new_profit)
state.set_water(water)
matrix = new_matrix
# save found data
found_profit_per_run.append(state.get_total_value())
time_needed.append(time.time() - start_run)
# One step is completed
runtimes += 1
# save final profit
final_profit = state.get_total_value()
# done
end_run = time.time()
# make dictionary to return data
return_values = {
"matrix": matrix,
"profit_per_run": found_profit_per_run,
"time_needed": time_needed,
"start_time": start_run,
"end_time": end_run,
"final_profit": final_profit
}
return return_values
def simulated_annealing_vrijstand(houses_to_place):
'''
Finds max vrijstand using Simulated annealing. Returns a dictionary with saved data.
Takes 2 arguments.
Houses_to_place: 1 for 20 houses, 2 for 40, 3 for 60.
'''
print("simulated annealing")
# save profit and time
found_profit_per_run = []
time_needed = []
# 90% change to move house, 10% change to move water
water_or_house = 0.9
# place houses
not_placed = True
# start placing
while (not_placed):
try:
matrix, houselist, water = init.initialize_matrix(
houses_to_place, WIDTH, HEIGTH)
not_placed = False
except:
not_placed = True
# calculate initial profit
fs.calculate_new_vrijstand_to_class(matrix, houselist)
result = prof.calculate_free_space(houselist)
found_profit_per_run.append(result)
# save findings
state = ss.saved_state(result, houselist, water)
# Start Temp Simulated Annealing
temperature = 50
start_temp = temperature
count = 0.0
# do the simulated annealing
start_run = time.time()
while (temperature > 0.08):
# update
if (count % 100 == 0):
print(count, temperature)
# make a copy of houselist
houselist = state.get_houselist_copy()
water = state.get_water_copy()
# random move water or house
if (water_or_house > random.random()):
house = random.randint(0, len(houselist) - 1)
moved, new_matrix = move.move_house(matrix, houselist[house], 100,
100)
else:
water_pool = random.randint(0, len(water.get_pools()) - 1)
moved, new_matrix = move.move_water(matrix,
water.get_pool(water_pool), 10,
10)
# if a house or water has succesfully been moved.
if (moved):
# calculate new profit
fs.calculate_new_vrijstand_to_class(new_matrix, houselist)
new_profit = prof.calculate_free_space(houselist)
# if more accepts
if (new_profit >= state.get_total_value()):
state.set_houselist(houselist)
state.set_total_value(new_profit)
state.set_water(water)
matrix = new_matrix
# else random accept depending on temperature
elif ((state.get_total_value() - new_profit) / temperature <
random.random()):
state.set_houselist(houselist)
state.set_total_value(new_profit)
state.set_water(water)
matrix = new_matrix
# decrease temperature
count += 1
temperature = start_temp * ((0.999)**abs(count))
# save findings
found_profit_per_run.append(state.get_total_value())
time_needed.append(time.time() - start_run)
# final profit
final_profit = state.get_total_value()
# done
end_run = time.time()
return_values = {
"matrix": matrix,
"profit_per_run": found_profit_per_run,
"time_needed": time_needed,
"start_time": start_run,
"end_time": end_run,
"final_profit": final_profit
}
return return_values