numbers_solver.py

from __future__ import division
from itertools import permutations

def solve(nums, target):
    ops = [("+", lambda x,y: x+y),
           ("/", lambda x,y: x/y),
           ("*", lambda x,y: x*y),
           ("-", lambda x,y: x-y)]
    closest = 1000
    curr_op = ""
    best_solution = []

    # Update closest based on numbers in the list, we might be done already.
    for n in nums:
        closest = min(closest, abs(target - n))
        if closest == 0:
            return (closest, best_solution)

    # Remove any zeros, they are useless.
    nums = filter(lambda x: x != 0, nums)

    # Choose two numbers to combine, then recursively call in for each possible
    # operation.
    for x, y in permutations(nums, 2):
        new_nums = nums[:]
        new_nums.remove(x)
        new_nums.remove(y)
        for op in ops:
            new_val = op[1](x,y)
            if int(new_val) != new_val:
                continue
            curr_op = "{0} {1} {2} = {3}".format(x, op[0], y, new_val)
            (score, soln) = solve(new_nums + [op[1](x,y)], target)
            if score < closest:
                closest = score
                best_solution = [curr_op] + soln
            if closest == 0:
                return (closest, best_solution)

    return (closest, best_solution)

if __name__ == "__main__":
    closest, soln = solve([75, 50, 25, 100, 5, 6], 893)
    for l in soln:
        print l