def search(problem, step, treshold):
time = 0
current_node = Node(problem.get_initial_state())
while True:
T = cooling_rate(time)
if T < treshold:
return current_node
adjacents = current_node.get_adjacents(step)
next = random.choice(adjacents)
value_difference = problem.find_value(next.state) - problem.find_value(
current_node.state)
if value_difference > 0:
current_node = next
else:
prob = math.exp(value_difference / T)
if random.random() <= prob:
current_node = next
time += 1
