模拟退火算法实现实例

import math
import random

# 定义目标函数（示例为一个简单的二维函数）
def cost_function(x, y):
    return (x - 2) ** 2 + (y + 3) ** 2

# 定义模拟退火算法
def simulated_annealing(cost_func, x_init, y_init, temp_init, temp_final, alpha, max_iter):
    
    # 初始化当前解和当前解对应的成本
    x_curr = x_init
    y_curr = y_init
    cost_curr = cost_func(x_curr, y_curr)
    
    # 初始化最优解和最优解对应的成本
    x_best = x_curr
    y_best = y_curr
    cost_best = cost_curr
    
    # 初始化温度
    temp_curr = temp_init
    
    # 迭代搜索
    for _ in range(max_iter):
        
        # 生成新解
        x_new = x_curr + random.uniform(-1, 1)
        y_new = y_curr + random.uniform(-1, 1)
        
        # 边界处理（根据实际问题进行修改）
        x_new = max(min(x_new, 10), -10)
        y_new = max(min(y_new, 10), -10)
        
        # 计算新解的成本
        cost_new = cost_func(x_new, y_new)
        
        # 计算成本变化
        cost_delta = cost_new - cost_curr
        
        # 判断是否接受新解
        if cost_delta < 0 or math.exp(-cost_delta / temp_curr) > random.random():
            x_curr = x_new
            y_curr = y_new
            cost_curr = cost_new
            
            # 更新最优解
            if cost_curr < cost_best:
                x_best = x_curr
                y_best = y_curr
                cost_best = cost_curr
        
        # 降低温度
        temp_curr *= alpha
    
    return x_best, y_best, cost_best

# 示例运行
x_init = random.uniform(-10, 10)
y_init = random.uniform(-10, 10)
temp_init = 100
temp_final = 0.1
alpha = 0.95
max_iter = 1000

x_best, y_best, cost_best = simulated_annealing(cost_function, x_init, y_init, temp_init, temp_final, alpha, max_iter)

print("最优解：", x_best, y_best)
print("最优解对应的成本：", cost_best)

本文仅为学习记录，如有错误欢迎指出。