帮我用PYTHON写一个遗传退火算法

当然可以，以下是一个简单的遗传退火算法的 Python 代码：

import random
import math

def simulated_annealing(cost_function, initial_solution, temperature, cooling_rate):
    current_solution = initial_solution
    best_solution = initial_solution
    while temperature > 1e-10:
        candidate_solution = current_solution[:]
        index = random.randint(0, len(candidate_solution) - 1)
        candidate_solution[index] += random.uniform(-1, 1)
        cost_diff = cost_function(candidate_solution) - cost_function(current_solution)
        if cost_diff < 0:
            current_solution = candidate_solution
            if cost_function(current_solution) < cost_function(best_solution):
                best_solution = current_solution
        else:
            if random.random() < math.exp(-cost_diff / temperature):
                current_solution = candidate_solution
        temperature *= cooling_rate
    return best_solution

# Example usage
def cost_function(x):
    return sum([xi ** 2 for xi in x])

initial_solution = [random.uniform(-10, 10) for _ in range(10)]
best_solution = simulated_annealing(cost_function, initial_solution, 100, 0.95)
print(best_solution)

这个代码实现了一个简单的遗传退火算法，用于求解一个多元函数的最小值。在这个例子中，我们使用了一个简单的二次函数作为 cost function，但是你可以根据自己的需求来定义自己的 cost function。