https://v2ex.com/t/895464

v2_895464.py

l1 = [23.17,3.2,1.22,0.32]
l2 = [7.36,4.16,3.20,1.69,1.28,1.28,0.96,0.96,0.90,0.64,0.64,0.64,0.50,0.50,0.32,0.32,0.32,0.32,0.32,0.32,0.32,0.32,0.32,0.32]

'''
Convert input to ints because of accuracy issue with float arithmetic.
'''
l1 = [round(x * 100) for x in l1]
l2 = [round(x * 100) for x in l2]

def group(l1, l2):
    def backtrack(i, ans):
        if i >= len(l2):
            return ans[:]
        j = 0
        while j < len(l1):
            if l1[j] >= l2[i]:
                ans.append(j)
                l1[j] -= l2[i]
                a = backtrack(i + 1, ans)
                l1[j] += l2[i]
                ans.pop()
                if a:
                    return a
            j += 1
        return []

    ret = {x : [] for x in l1}

    if sum(l1) != sum(l2):
        # No Answer
        return ret

    ans = backtrack(0, [])

    for i, x in enumerate(ans):
        ret[l1[x]].append(l2[i])

    return ret

print("l1 =", l1)
print("l2 =", l2)
print("ans =", group(l1, l2))

'''
l1 = [2317, 320, 122, 32]
l2 = [736, 416, 320, 169, 128, 128, 96, 96, 90, 64, 64, 64, 50, 50, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32]
ans = {2317: [736, 416, 320, 169, 128, 128, 96, 96, 64, 64, 50, 50], 320: [64, 32, 32, 32, 32, 32, 32, 32, 32], 122: [90, 32], 32: [32]}
'''

This brute force approach will not work for reasonably larger dataset. We will need to apply further heuristics to prune the search space.