Python算法教程：背包问题

背包问题可用这一套术语来定义:有一组想要带在身边的物品，每个物品都有各自的质量和价值。但问题是背包有一个最大容量，如何才能使其中所带物品的总价值达到最高呢？

选取一些元素组成一个带权值的有限子集，使其权值最大化。

无限制背包问题

当作一个模式来看待：以某种形式来定义问题的子问题，递归这些子问题之间的关系，然后确保每个子问题都只被计算一次。

围绕这背包容量中最后一个单元是否有用来进行。

def unbound_knapsack(weights, values, capacity):
    res = [0]
    for cu in range(1, capacity+1):
        current = res[cu-1]
        for wi, wu in enumerate(weights):
            if wu > cu:
                continue
            current = max(current, values[wi]+res[cu-wu])
        res.append(current)
    return res[-1]


weights = [20, 10, 8]
values = [19, 31, 19]
capacity = 20
print(unbound_knapsack(weights, values, capacity))  # 62

0-1背包问题

每个对象最多只能用一次。

是否要纳入最后一个对象

def knapsack(weights, values, capacity):
    n = len(weights)
    res = [[0]*(capacity+1) for _ in range(n+1)]
    flag = [[False]*(capacity+1) for _ in range(n+1)]
    for nu in range(1, n+1):
        for cu in range(1, capacity+1):
            res[nu][cu] = pre = res[nu-1][cu]
            if weights[nu-1] > cu:
                continue
            current = values[nu-1] + res[nu-1][cu-weights[nu-1]]
            res[nu][cu] = max(pre, current)
            flag[nu][cu] = current > pre
    return res, flag


weights = [20, 10, 8]
values = [19, 31, 19]
capacity = 20

res, flag = knapsack(weights, values, capacity)
print(res[-1][-1])  # 50

wi, cur, items = len(weights), capacity, set()
while wi > 0 and cur > 0:
    if flag[wi][cur]:
        items.add(wi-1)
        cur -= weights[wi-1]
    wi -= 1
thing = {weights[i]: values[i] for i in items}
print(thing)    # {10: 31, 8: 19}

（最近更新：2019年09月06日）