动态规划之钢条切割问题——Rod-cutting problem

本文链接：https://blog.csdn.net/liuxiang15/article/details/87174771

问题描述

问题的最优解包含相关子问题的最优解，这些子问题我们可以独立求解。

应用以下公式：

方法1：暴力法

长度为n的钢条可以有2^(n-1)中切割方案,因为离钢条左端距离分别为1，2，……，（n-1）的位置都有切或不切（0或1）两种选择，

但是我认为有好多重复的情况，假设切割的位置存储在数组中，那么[1]和[n-1]其实是一样的，因为都代表切成长为1和(n-1)的两段。但是具体重复情况（类似回文数的数量）有多少我一下子不知道怎么计算，但是数量级还是指数级的！！！

以下的图中伪代码基本相当于python代码，重大区别就是数组下标 i 有些差别！！！在我写的代码中

length = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
profit = [1, 5, 8, 9, 10, 17, 17, 20, 24, 30]

def cut_rod(p, n):
    if n == 0:
        return 0
    q = -1
    for i in range(n):
        q = max(q, p[i]+cut_rod(p, n-i))
    return q

print(cut_rod(profit, 9))

运行结果：

C:\Users\liuxiang15\Desktop\homework3>python rod_cutting.py
Traceback (most recent call last):
  File "rod_cutting.py", line 11, in <module>
    print(cut_rod(profit, 9))
  File "rod_cutting.py", line 9, in cut_rod
    q = max(q, p[i]+cut_rod(p, n-i))
  File "rod_cutting.py", line 9, in cut_rod
    q = max(q, p[i]+cut_rod(p, n-i))
  File "rod_cutting.py", line 9, in cut_rod
    q = max(q, p[i]+cut_rod(p, n-i))
  [Previous line repeated 993 more times]
  File "rod_cutting.py", line 8, in cut_rod
    for i in range(n):
RecursionError: maximum recursion depth exceeded in comparison

方法2：带备忘的自顶向下法

为了避免“自顶向下递归算法”中出现的重复计算重叠子问题的情况，用数组r存储不同长度下的最大效益

代码：

length = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
profit = [1, 5, 8, 9, 10, 17, 17, 20, 24, 30]

def memorized_cut_rod(p, n):
    r = [-1]*(n+1)
    return memorized_cut_rod_aux(p, n, r)

def memorized_cut_rod_aux(p, n, r):
    if r[n] > 0:
        return r[n]
    q = -1
    if n == 0:
        q = 0
    else:
        for i in range(1, n+1):
            q = max(q, p[i-1]+memorized_cut_rod_aux(p, n-i, r))
    r[n] = q
    return q

for i in range(1, 11):
    print(memorized_cut_rod(profit, i))

运行结果：

C:\Users\liuxiang15\Desktop\homework3>python rod_cutting.py
1
5
8
10
13
17
18
22
25
30

方法3：自底向上法

length = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
profit = [1, 5, 8, 9, 10, 17, 17, 20, 24, 30]

def bottom_up_cut_rod(p, n):
    r = [0]*(n+1)
    for j in range(1, n+1):
        q = -1
        for i in range(1,j+1):
            q = max(q, p[i-1]+r[j-i])
        r[j] = q
    return r[n]  

for i in range(1, 11):
    print(bottom_up_cut_rod(profit, i))

C:\Users\liuxiang15\Desktop\homework3>python rod_cutting.py
1
5
8
10
13
17
18
22
25
30

方法4：扩展的自底向上法(s数组存储切割位置)

代码增加了一个功能，待切割长度n可以大于len(profit)

length = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
profit = [1, 5, 8, 9, 10, 17, 17, 20, 24, 30]

def extended_bottom_up_cut_rod(p, n):
    r = [0]*(n+1)
    s = [0]*(n+1)
    for j in range(1, n+1):
        q = -1
        for i in range(1,j+1):
            if i <= len(profit):
                if q < p[i-1]+r[j-i]:
                    q = p[i-1]+r[j-i]
                    s[j] = i
            else:
                if q < r[i]+r[j-i]:
                    q = r[i] + r[j-i]
                    s[j] = i
        r[j] = q
    # return r[n],s[n]
    return r,s
print(extended_bottom_up_cut_rod(profit, 15))

C:\Users\liuxiang15\Desktop\homework3>python rod_cutting.py
([0, 1, 5, 8, 10, 13, 17, 18, 22, 25, 30, 31, 35, 38, 40, 43], [0, 1, 2, 3, 2, 2, 6, 1, 2, 3, 10, 1, 2, 3, 2, 2])

方法4.1：使用方法4函数来存储切割段长在数组中

# method of how to cut the rod
def cut_rod_solution(p, n):
    (r,s) = extended_bottom_up_cut_rod(p, n)
    print(r)
    print(s)
    cut_method = []
    while n > 0:
        cut_method.append(s[n])
        # print(s[n])
        n -= s[n]
    return cut_method

print(cut_rod_solution(profit, 15))

C:\Users\liuxiang15\Desktop\homework3>python rod_cutting.py
[0, 1, 5, 8, 10, 13, 17, 18, 22, 25, 30, 31, 35, 38, 40, 43]
[0, 1, 2, 3, 2, 2, 6, 1, 2, 3, 10, 1, 2, 3, 2, 2]
[2, 3, 10]