手术排程问题：已知各手术术式平均用时，每日有140台手术需要安排到20个手术间，最大化手术室黄金时间利用率并且最小化医院20个手术间用时标准差的启发式算法python具体代码

这是一个NP难问题，可以采用启发式算法来逼近最优解。以下是一个基于模拟退火的启发式算法的Python代码：

import random
import math
import copy

# 手术间数量
ROOM_NUM = 20

# 每天手术数量
DAILY_OP_NUM = 140

# 每个手术室每天最大使用时间
MAX_TIME_PER_ROOM = 480

# 每个手术的平均用时
OP_TIME = [60, 90, 120, 150, 180]

# 初始温度
INIT_TEMP = 100

# 降温速率
COOL_RATE = 0.99

# 终止温度
END_TEMP = 1

# 迭代次数
ITER_NUM = 500

# 模拟退火算法
def simulated_annealing():
    # 随机生成初始解
    solution = [[[] for j in range(ROOM_NUM)] for i in range(DAILY_OP_NUM)]
    for i in range(DAILY_OP_NUM):
        for j in range(ROOM_NUM):
            op_time = random.choice(OP_TIME)
            solution[i][j] = op_time
    # 计算初始解的黄金时间利用率和用时标准差
    max_time_per_day = MAX_TIME_PER_ROOM * ROOM_NUM
    time_used = [0 for i in range(DAILY_OP_NUM)]
    for i in range(DAILY_OP_NUM):
        for j in range(ROOM_NUM):
            time_used[i] += solution[i][j]
    time_used.sort()
    gold_time_used = sum(time_used[-DAILY_OP_NUM//2:])
    time_std = math.sqrt(sum([(t - gold_time_used)**2 for t in time_used]) / DAILY_OP_NUM)
    # 初始化温度
    temp = INIT_TEMP
    # 迭代降温
    for i in range(ITER_NUM):
        # 随机选择一个手术进行调整
        op_idx = random.randint(0, DAILY_OP_NUM-1)
        from_room_idx = random.randint(0, ROOM_NUM-1)
        to_room_idx = random.randint(0, ROOM_NUM-1)
        while from_room_idx == to_room_idx:
            to_room_idx = random.randint(0, ROOM_NUM-1)
        # 复制当前解
        new_solution = copy.deepcopy(solution)
        # 调整手术
        new_solution[op_idx][from_room_idx], new_solution[op_idx][to_room_idx] = new_solution[op_idx][to_room_idx], new_solution[op_idx][from_room_idx]
        # 计算新解的黄金时间利用率和用时标准差
        new_time_used = [0 for i in range(DAILY_OP_NUM)]
        for i in range(DAILY_OP_NUM):
            for j in range(ROOM_NUM):
                new_time_used[i] += new_solution[i][j]
        new_time_used.sort()
        new_gold_time_used = sum(new_time_used[-DAILY_OP_NUM//2:])
        new_time_std = math.sqrt(sum([(t - new_gold_time_used)**2 for t in new_time_used]) / DAILY_OP_NUM)
        # 计算能量差
        delta_e = new_gold_time_used - gold_time_used + (time_std - new_time_std) / 10
        # 判断是否接受新解
        if delta_e < 0 or math.exp(-delta_e/temp) > random.random():
            solution = new_solution
            gold_time_used = new_gold_time_used
            time_std = new_time_std
        # 降温
        temp *= COOL_RATE
        if temp < END_TEMP:
            break
    return solution, gold_time_used, time_std

# 测试
solution, gold_time_used, time_std = simulated_annealing()
print('黄金时间利用率：', gold_time_used / (MAX_TIME_PER_ROOM * ROOM_NUM * DAILY_OP_NUM))
print('用时标准差：', time_std)
print('手术安排方案：')
for i in range(DAILY_OP_NUM):
    print(solution[i])

其中，黄金时间利用率是指每天手术间使用时间的中位数占总时间的比例，用时标准差是用来衡量手术间使用时间分布的均匀性的指标。模拟退火算法的基本思路是随机生成初始解，然后迭代地在解空间进行随机搜索，通过接受劣解的概率逐渐降低温度，最终得到一个相对较优的解。