Related
Input: n arrays of integers of length p.
Output: An array of p integers built by copying contiguous subarrays of the input arrays into matching indices of the output, satisfying the following conditions.
At most one subarray is used from each input array.
Every index of the output array is filled from exactly one subarray.
The output array has the minimum possible sum.
Suppose I have 2 arrays:
[1,7,2]
[2,1,8]
So if I choose a subarray [1,7] from array 1 and subarray [8] from array 2. since these 2 subarrays are not overlapping for any index and are contiguous. We are also not taking any subarray twice from an array from which we have already chosen a subarray.
We have the number of elements in the arrays inside the collection = 2 + 1 = 3, which is the same as the length of the individual array (i.e. len(array 1) which is equal to 3). So, this collection is valid.
The sum here for [1,7] and [8] is 1 + 7 + 8 = 16
We have to find a collection of such subarrays such that the total sum of the elements of subarrays is minimum.
A solution to the above 2 arrays would be a collection [2,1] from array 1 and [2] from array 2.
This is a valid collection and the sum is 2 + 1 + 2 = 5 which is the minimum sum for any such collection in this case.
I cannot think of any optimal or correct approach, so I need help.
Some Ideas:
I tried a greedy approach by choosing minimum elements from all array for a particular index since the index is always increasing (non-overlapping) after a valid choice, I don't have to bother about storing minimum value indices for every array. But this approach is clearly not correct since it will visit the same array twice.
Another method I thought was to start from the 0th index for all arrays and start storing their sum up to k elements for every array since the no. of arrays are finite, I can store the sum upto k elements in an array. Now I tried to take a minimum across these sums and for a "minimum sum", the corresponding subarray giving this sum (i.e. k such elements in that array) can be a candidate for a valid subarray of size k, thus if we take this subarray, we can add a k + 1-th element corresponding to every array into their corresponding sum and if the original minimum still holds, then we can keep on repeating this step. When the minima fail, we can consider the subarray up to the index for which minima holds and this will be a valid starting subarray. However, this approach will also clearly fail because there could exist another subarray of size < k giving minima along with remaining index elements from our subarray of size k.
Sorting is not possible either, since if we sort then we are breaking consecutive condition.
Of course, there is a brute force method too.
I am thinking, working through a greedy approach might give a progress in the approach.
I have searched on other Stackoverflow posts, but couldn't find anything which could help my problem.
To get you started, here's a recursive branch-&-bound backtracking - and potentially exhaustive - search. Ordering heuristics can have a huge effect on how efficient these are, but without mounds of "real life" data to test against there's scant basis for picking one over another. This incorporates what may be the single most obvious ordering rule.
Because it's a work in progress, it prints stuff as it goes along: all solutions found, whenever they meet or beat the current best; and the index at which a search is cut off early, when that happens (because it becomes obvious that the partial solution at that point can't be extended to meet or beat the best full solution known so far).
For example,
>>> crunch([[5, 6, 7], [8, 0, 3], [2, 8, 7], [8, 2, 3]])
displays
new best
L2[0:1] = [2] 2
L1[1:2] = [0] 2
L3[2:3] = [3] 5
sum 5
cut at 2
L2[0:1] = [2] 2
L1[1:3] = [0, 3] 5
sum 5
cut at 2
cut at 2
cut at 2
cut at 1
cut at 1
cut at 2
cut at 2
cut at 2
cut at 1
cut at 1
cut at 1
cut at 0
cut at 0
So it found two ways to get a minimal sum 5, and the simple ordering heuristic was effective enough that all other paths to full solutions were cut off early.
def disp(lists, ixs):
from itertools import groupby
total = 0
i = 0
for k, g in groupby(ixs):
j = i + len(list(g))
chunk = lists[k][i:j]
total += sum(chunk)
print(f"L{k}[{i}:{j}] = {chunk} {total}")
i = j
def crunch(lists):
n = len(lists[0])
assert all(len(L) == n for L in lists)
# Start with a sum we know can be beat.
smallest_sum = sum(lists[0]) + 1
smallest_ixs = [None] * n
ixsofar = [None] * n
def inner(i, sumsofar, freelists):
nonlocal smallest_sum
assert sumsofar <= smallest_sum
if i == n:
print()
if sumsofar < smallest_sum:
smallest_sum = sumsofar
smallest_ixs[:] = ixsofar
print("new best")
disp(lists, ixsofar)
print("sum", sumsofar)
return
# Simple greedy heuristic: try available lists in the order
# of smallest-to-largest at index i.
for lix in sorted(freelists, key=lambda lix: lists[lix][i]):
L = lists[lix]
newsum = sumsofar
freelists.remove(lix)
# Try all slices in L starting at i.
for j in range(i, n):
newsum += L[j]
# ">" to find all smallest answers;
# ">=" to find just one (potentially faster)
if newsum > smallest_sum:
print("cut at", j)
break
ixsofar[j] = lix
inner(j + 1, newsum, freelists)
freelists.add(lix)
inner(0, 0, set(range(len(lists))))
How bad is brute force?
Bad. A brute force way to compute it: say there are n lists each with p elements. The code's ixsofar vector contains p integers each in range(n). The only constraint is that all occurrences of any integer that appears in it must be consecutive. So a brute force way to compute the total number of such vectors is to generate all p-tuples and count the number that meet the constraints. This is woefully inefficient, taking O(n**p) time, but is really easy, so hard to get wrong:
def countb(n, p):
from itertools import product, groupby
result = 0
seen = set()
for t in product(range(n), repeat=p):
seen.clear()
for k, g in groupby(t):
if k in seen:
break
seen.add(k)
else:
#print(t)
result += 1
return result
For small arguments, we can use that as a sanity check on the next function, which is efficient. This builds on common "stars and bars" combinatorial arguments to deduce the result:
def count(n, p):
# n lists of length p
# for r regions, r from 1 through min(p, n)
# number of ways to split up: comb((p - r) + r - 1, r - 1)
# for each, ff(n, r) ways to spray in list indices = comb(n, r) * r!
from math import comb, prod
total = 0
for r in range(1, min(n, p) + 1):
total += comb(p-1, r-1) * prod(range(n, n-r, -1))
return total
Faster
Following is the best code I have for this so far. It builds in more "smarts" to the code I posted before. In one sense, it's very effective. For example, for randomized p = n = 20 inputs it usually finishes within a second. That's nothing to sneeze at, since:
>>> count(20, 20)
1399496554158060983080
>>> _.bit_length()
71
That is, trying every possible way would effectively take forever. The number of cases to try doesn't even fit in a 64-bit int.
On the other hand, boost n (the number of lists) to 30, and it can take an hour. At 50, I haven't seen a non-contrived case finish yet, even if left to run overnight. The combinatorial explosion eventually becomes overwhelming.
OTOH, I'm looking for the smallest sum, period. If you needed to solve problems like this in real life, you'd either need a much smarter approach, or settle for iterative approximation algorithms.
Note: this is still a work in progress, so isn't polished, and prints some stuff as it goes along. Mostly that's been reduced to running a "watchdog" thread that wakes up every 10 minutes to show the current state of the ixsofar vector.
def crunch(lists):
import datetime
now = datetime.datetime.now
start = now()
n = len(lists[0])
assert all(len(L) == n for L in lists)
# Start with a sum we know can be beat.
smallest_sum = min(map(sum, lists)) + 1
smallest_ixs = [None] * n
ixsofar = [None] * n
import threading
def watcher(stop):
if stop.wait(60):
return
lix = ixsofar[:]
while not stop.wait(timeout=600):
print("watch", now() - start, smallest_sum)
nlix = ixsofar[:]
for i, (a, b) in enumerate(zip(lix, nlix)):
if a != b:
nlix.insert(i,"--- " + str(i) + " -->")
print(nlix)
del nlix[i]
break
lix = nlix
stop = threading.Event()
w = threading.Thread(target=watcher, args=[stop])
w.start()
def inner(i, sumsofar, freelists):
nonlocal smallest_sum
assert sumsofar <= smallest_sum
if i == n:
print()
if sumsofar < smallest_sum:
smallest_sum = sumsofar
smallest_ixs[:] = ixsofar
print("new best")
disp(lists, ixsofar)
print("sum", sumsofar, now() - start)
return
# If only one input list is still free, we have to take all
# of its tail. This code block isn't necessary, but gives a
# minor speedup (skips layers of do-nothing calls),
# especially when the length of the lists is greater than
# the number of lists.
if len(freelists) == 1:
lix = freelists.pop()
L = lists[lix]
for j in range(i, n):
ixsofar[j] = lix
sumsofar += L[j]
if sumsofar >= smallest_sum:
break
else:
inner(n, sumsofar, freelists)
freelists.add(lix)
return
# Peek ahead. The smallest completion we could possibly get
# would come from picking the smallest element in each
# remaining column (restricted to the lists - rows - still
# available). This probably isn't achievable, but is an
# absolute lower bound on what's possible, so can be used to
# cut off searches early.
newsum = sumsofar
for j in range(i, n): # pick smallest from column j
newsum += min(lists[lix][j] for lix in freelists)
if newsum >= smallest_sum:
return
# Simple greedy heuristic: try available lists in the order
# of smallest-to-largest at index i.
sortedlix = sorted(freelists, key=lambda lix: lists[lix][i])
# What's the next int in the previous slice? As soon as we
# hit an int at least that large, we can do at least as well
# by just returning, to let the caller extend the previous
# slice instead.
if i:
prev = lists[ixsofar[i-1]][i]
else:
prev = lists[sortedlix[-1]][i] + 1
for lix in sortedlix:
L = lists[lix]
if prev <= L[i]:
return
freelists.remove(lix)
newsum = sumsofar
# Try all non-empty slices in L starting at i.
for j in range(i, n):
newsum += L[j]
if newsum >= smallest_sum:
break
ixsofar[j] = lix
inner(j + 1, newsum, freelists)
freelists.add(lix)
inner(0, 0, set(range(len(lists))))
stop.set()
w.join()
Bounded by DP
I've had a lot of fun with this :-) Here's the approach they were probably looking for, using dynamic programming (DP). I have several programs that run faster in "smallish" cases, but none that can really compete on a non-contrived 20x50 case. The runtime is O(2**n * n**2 * p). Yes, that's more than exponential in n! But it's still a minuscule fraction of what brute force can require (see above), and is a hard upper bound.
Note: this is just a loop nest slinging machine-size integers, and using no "fancy" Python features. It would be easy to recode in C, where it would run much faster. As is, this code runs over 10x faster under PyPy (as opposed to the standard CPython interpreter).
Key insight: suppose we're going left to right, have reached column j, the last list we picked from was D, and before that we picked columns from lists A, B, and C. How can we proceed? Well, we can pick the next column from D too, and the "used" set {A, B, C} doesn't change. Or we can pick some other list E, the "used" set changes to {A, B, C, D}, and E becomes the last list we picked from.
Now in all these cases, the details of how we reached state "used set {A, B, C} with last list D at column j" make no difference to the collection of possible completions. It doesn't matter how many columns we picked from each, or the order in which A, B, C were used: all that matters to future choices is that A, B, and C can't be used again, and D can be but - if so - must be used immediately.
Since all ways of reaching this state have the same possible completions, the cheapest full solution must have the cheapest way of reaching this state.
So we just go left to right, one column at a time, and remember for each state in the column the smallest sum reaching that state.
This isn't cheap, but it's finite ;-) Since states are subsets of row indices, combined with (the index of) the last list used, there are 2**n * n possible states to keep track of. In fact, there are only half that, since the way sketched above never includes the index of the last-used list in the used set, but catering to that would probably cost more than it saves.
As is, states here are not represented explicitly. Instead there's just a large list of sums-so-far, of length 2**n * n. The state is implied by the list index: index i represents the state where:
i >> n is the index of the last-used list.
The last n bits of i are a bitset, where bit 2**j is set if and only if list index j is in the set of used list indices.
You could, e.g., represent these by dicts mapping (frozenset, index) pairs to sums instead, but then memory use explodes, runtime zooms, and PyPy becomes much less effective at speeding it.
Sad but true: like most DP algorithms, this finds "the best" answer but retains scant memory of how it was reached. Adding code to allow for that is harder than what's here, and typically explodes memory requirements. Probably easiest here: write new to disk at the end of each outer-loop iteration, one file per column. Then memory use isn't affected. When it's done, those files can be read back in again, in reverse order, and mildly tedious code can reconstruct the path it must have taken to reach the winning state, working backwards one column at a time from the end.
def dumbdp(lists):
import datetime
_min = min
now = datetime.datetime.now
start = now()
n = len(lists)
p = len(lists[0])
assert all(len(L) == p for L in lists)
rangen = range(n)
USEDMASK = (1 << n) - 1
HUGE = sum(sum(L) for L in lists) + 1
new = [HUGE] * (2**n * n)
for i in rangen:
new[i << n] = lists[i][0]
for j in range(1, p):
print("working on", j, now() - start)
old = new
new = [HUGE] * (2**n * n)
for key, g in enumerate(old):
if g == HUGE:
continue
i = key >> n
new[key] = _min(new[key], g + lists[i][j])
newused = (key & USEDMASK) | (1 << i)
for i in rangen:
mask = 1 << i
if newused & mask == 0:
newkey = newused | (i << n)
new[newkey] = _min(new[newkey],
g + lists[i][j])
result = min(new)
print("DONE", result, now() - start)
return result
I have been asked the following question in an interview and I am still thinking of an efficient way of doing it.
You have an array whose numbers represent percentages of liquids in a barrel.You also have an API with a method: combine(int x,int y) which takes two input percentages in the array and combines the liquid from one barrel to another. By using this information you have to find maximum number of barrels that can be possible with 100% liquid.
Example 1. Array: 10,15,20,35,55,65
Ans:Number of barrels would be 2.
Since combine(65,35)---one 100% barrel,
combine(55,20)--75% barrel, next combine(75,15)--90% next combine(90,10)--100%--1 barrel
So total 2 barrels
Example 2: 99,99,99
Ans: Number of barrels would be 1 here since you do combine(99,99)--you get one 100% barrel the rest of the liquid is wasted and you can't combine any other barrel with the third 99% barrel to make it 100
Note:once you pour liquid from one barrel to another you can't use it again
for ex: combine(55,15)--70% barrel. You can use 70% barrel but not 55% and 15% barrels.
You may indeed look at algorithms for bin packing problem. This student paper indicates four of them. The one with best approximation is Decreasing First Fit algo.
It comes down to quick sort (in place, O(nlogn) time complexity in average and O(n2) in worst case), and then First Fit.
First Fit comes down to scan the bins in order, and place the new item in the first bin that is large enough to hold it. If there is no bin into which current object fits, start a new bin.
To FF in O(nlogn) complexity, use a Max Winner Tree data structure. It has n external nodes (players) and n-1 internal nodes (winner for each match),
and the winner is the player with max value.
Assuming all percentages in the given array are less than 100 (in any case if there were elements at or greater than 100, we could count and remove them immediately), each barrel of 100% cannot be created from less than two array elements, and the number of 100% barrels cannot be more than the sum of the array divided by 100. Therefore, the possibilities to examine are bound by:
maxNumBarrels array = min (div (sum array) 100) (div (length array) 2)
The following Haskell code provides the function, divide, which partitions the array into all variations of n partitions without repetition (that is, both partition order and element order within partitions are ignored). The function, maxBarrels, searches backwards, dividing the array first into maxNumBarrels partitions (searching for results with maxNumBarrels elements who's sum is >=100), and then progressively smaller numbers of partitions until an answer is found or a null set is returned.
import Data.Map (adjust, fromList, toList)
divide xs n = divide' xs (zip [0..] (replicate n [])) where
divide' [] result = [result]
divide' (x:xs) result = do
index <- indexes
divide' xs (toList $ adjust (x :) index (fromList result)) where
populated = map fst . filter (not . null . snd) $ result
indexes = populated ++ if any (null . snd) result
then [length populated]
else []
maxBarrels xs = allDivided maxNumBarrels where
maxNumBarrels = min (div (sum xs) 100) (div (length xs) 2)
allDivided count | count == 0 = []
| not (null divided) = divided
| otherwise = allDivided (count - 1)
where divided = filter ((==count) . length)
. map (filter ((>=100) . sum))
. map (map snd)
. divide xs $ count
OUTPUT:
*Main> maxBarrels [10,15,20,35,55,65]
[[[55,20,15,10],[65,35]],[[55,35,10],[65,20,15]]]
*Main> maxBarrels [99,99,99]
[[[99,99,99]]]
*Main> maxBarrels [99,99,99,10,15,25,35,55,65]
[[[15,10,99],[25,99],[35,99],[65,55]] ...(the first of 144 immediate results)...
Is there a more efficient approach to computing a histogram than a binary search for a non-linear bin distribution?
I'm actually only interested in the bit of the algorithm that matches the key (value) to the bin (the transfer function?) , i.e. for a bunch of floating point values I just want to know the appropriate bin index for each value.
I know that for a linear bin distribution you can get O(1) by dividing the value by the bin width, and that for non linear bins a binary search gets you O(logN). My current implementation uses a binary search on unequal bin widths.
In the spirit of improving efficiency I was curious as to whether you could use a hash function to map a value to its appropriate bin and achieve O(1) time complexity when you have bins of unequal widths?
In some simple cases you can get O(1).
Suppose, your values are 8-bit, from 0 to 255.
If you split them into 8 bins of sizes 2, 2, 4, 8, 16, 32, 64, 128, then the bin value ranges will be: 0-1, 2-3, 4-7, 8-15, 16-31, 32-63, 64-127, 128-255.
In binary these ranges look like:
0000000x (bin 0)
0000001x
000001xx
00001xxx
0001xxxx
001xxxxx
01xxxxxx
1xxxxxxx (bin 7)
So, if you can quickly (in O(1)) count how many most significant zero bits there are in the value, you can get the bin number from it.
In this particular case you may precalculate a look-up table of 256 elements, containing the bin number and finding the appropriate bin for a value is just one table look-up.
Actually, with 8-bit values you can use bins of arbitrary sizes since the look-up table is small.
If you were to go with bins of sizes of powers of 2, you could reuse this look-up table for 16-bit values as well. And you'd need two look-ups. You can extend it to even longer values.
Ordinary hash functions are intended to scatter different values quite randomly across some range. A single-bit difference in arguments may lead to dozens of bits different in results. For that reason, ordinary hash functions are not suitable for the situation described in the question.
An alternative is to build an array P with entries that index into the table B of bin limits. Given some value x, we find the bin j it belongs to (or sometimes a nearby bin) via j = P[⌊x·r⌋] where r is a ratio that depends on the size of P and the maximum value in B. The effectiveness of this approach depends on the values in B and the size of P.
The behavior of functions like P[⌊x·r⌋] can be seen via the python code shown below. (The method is about the same in any programming language. However, tips for Python-to-C are given below.) Suppose the code is stored in file histobins.py and loaded into the ipython interpreter with the command import histobins as hb. Then a command like hb.betterparts(27, 99, 9, 80,155) produces output like
At 80 parts, steps = 20 = 7+13
At 81 parts, steps = 16 = 7+9
At 86 parts, steps = 14 = 6+8
At 97 parts, steps = 13 = 12+1
At 108 parts, steps = 12 = 3+9
At 109 parts, steps = 12 = 8+4
At 118 parts, steps = 12 = 6+6
At 119 parts, steps = 10 = 7+3
At 122 parts, steps = 10 = 3+7
At 141 parts, steps = 10 = 5+5
At 142 parts, steps = 10 = 4+6
At 143 parts, steps = 9 = 7+2
These parameters to betterparts set nbins=27, topsize=99, seed=9, plo=80, phi=155 which creates a test set of 27 bins for values from 0 to 99, with random seed 9, and size of P from 80 to 155-1. The number of “steps” is the number of times the two while loops in testparts() operated during a test with 10*nbins values from 0 to topsize. Eg, “At 143 parts, steps = 9 = 7+2” means that when the size of P is 143, out of 270 trials, 261 times P[⌊x·r⌋] produced the correct index at once; 7 times the index had to be decreased, and twice it had to be increased.
The general idea of the method is to trade off space for time. Another tradeoff is preparation time versus operation time. If you are going to be doing billions of lookups, it is worthwhile to do a few thousand trials to find a good value of |P|, the size of P. If you are going to be doing only a few millions of lookups, it might be better to just pick some large value of |P| and run with it, or perhaps just run betterparts over a narrow range. Instead of doing 75 tests as above, if we start with larger |P| fewer tests may give a good enough result. For example, 10 tests via “hb.betterparts(27, 99, 9, 190,200)” produces
At 190 parts, steps = 11 = 5+6
At 191 parts, steps = 5 = 3+2
At 196 parts, steps = 5 = 4+1
As long as P fits into some level of cache (along with other relevant data) making |P| larger will speed up access. So, making |P| as large as practical is a good idea. As |P| gets larger, the difference in performance between one value of |P| and the next gets smaller and smaller. The limiting factors on speed then include time to multiply and time to set up while loops. One approach for faster multiplies may be to choose a power of 2 as a multiplier; compute |P| to match; then use shifts or adds to exponents instead of multiplies. One approach to spending less time setting up while loops is to move the statement if bins[bin] <= x < bins[bin+1]: (or its C equivalent, see below) to before the while statements and do the while's only if the if statement fails.
Python code is shown below. Note, in translating from Python to C,
• # begins a comment
• def begins a function
• a statement like ntest, right, wrong, x = 10*nbins, 0, 0, 0 assigns values to respective identifiers
• a statement like return (ntest, right, wrong, stepdown, stepup) returns a tuple of 5 values that the caller can assign to a tuple or to respective identifiers
• the scope of a def, while, or if ends with a line not indented farther than the def, while, or if
• bins = [0] initializes a list (an extendible indexable array) with value 0 as its initial entry
• bins.append(t) appends value t at the end of list bins
• for i,j in enumerate(p): runs a loop over the elements of iterable p (in this case, p is a list), making the index i and corresponding entry j == p[i] available inside the loop
• range(nparts) stands for a list of the values 0, 1, ... nparts-1
• range(plo, phi) stands for a list of the values plo, plo+1, ... phi-1
• if bins[bin] <= x < bins[bin+1] means if ((bins[bin] <= x) && (x < bins[bin+1]))
• int(round(x*float(nparts)/topsize))) actually rounds x·r, instead of computing ⌊x·r⌋ as advertised above
def makebins(nbins, topsize):
bins, t = [0], 0
for i in range(nbins):
t += random.random()
bins.append(t)
for i in range(nbins+1):
bins[i] *= topsize/t
bins.append(topsize+1)
return bins
#________________________________________________________________
def showbins(bins):
print ''.join('{:6.2f} '.format(x) for x in bins)
def showparts(nbins, bins, topsize, nparts, p):
ratio = float(topsize)/nparts
for i,j in enumerate(p):
print '{:3d}. {:3d} {:6.2f} {:7.2f} '.format(i, j, bins[j], i*ratio)
print 'nbins: {} topsize: {} nparts: {} ratio: {}'.format(nbins, topsize, nparts, ratio)
print 'p = ', p
print 'bins = ',
showbins(bins)
#________________________________________________________________
def testparts(nbins, topsize, nparts, seed):
# Make bins and make lookup table p
import random
if seed > 0: random.seed(seed)
bins = makebins(nbins,topsize)
ratio, j, p = float(topsize)/nparts, 0, range(nparts)
for i in range(nparts):
while j<nbins and i*ratio >= bins[j+1]:
j += 1
p[i] = j
p.append(j)
#showparts(nbins, bins, topsize, nparts, p)
# Count # of hits and steps with avg. of 10 items per bin
ntest, right, wrong, x = 10*nbins, 0, 0, 0
delta, stepdown, stepup = topsize/float(ntest), 0, 0
for i in range(ntest):
bin = p[min(nparts, max(0, int(round(x*float(nparts)/topsize))))]
while bin < nbins and x >= bins[bin+1]:
bin += 1; stepup += 1
while bin > 0 and x < bins[bin]:
bin -= 1; stepdown += 1
if bins[bin] <= x < bins[bin+1]: # Test if bin is correct
right += 1
else:
wrong += 1
print 'Wrong bin {} {:7.3f} at x={:7.3f} Too {}'.format(bin, bins[bin], x, 'high' if bins[bin] > x else 'low')
x += delta
return (ntest, right, wrong, stepdown, stepup)
#________________________________________________________________
def betterparts(nbins, topsize, seed, plo, phi):
beststep = 1e9
for parts in range(plo, phi):
ntest, right, wrong, stepdown, stepup = testparts(nbins, topsize, parts, seed)
if wrong: print 'Error with ', parts, ' parts'
steps = stepdown + stepup
if steps <= beststep:
beststep = steps
print 'At {:3d} parts, steps = {:d} = {:d}+{:d}'.format(parts, steps, stepdown, stepup)
#________________________________________________________________
Interpolation search is your friend. It's kind of an optimistic, predictive binary search where it guesses where the bin should be based on a linear assumption about the distribution of inputs, rather than just splitting the search space in half at each step. It will be O(1) if the linear assumption is true, but still works (though more slowly) when the assumption is not. To the degree that its predictions are accurate, the search is fast.
Depends on the implementation of the hashing and the type of data you're working with. For smaller data sets a more simple algorithm like binary search might outperform constant lookup if the lookup-overhead of hashing is larger on average.
The usual implementation of hashing, consists of an array of linked lists and a hashing function that maps a string to an index in the array of linked lists. There's a thing called the load factor, which is the number of elements in the hash map / length of the linked-list array. Thus for load factors < 1 you'll achieve constant lookup in the best case because no linked-list will contain more than one element (best case).
There's only one way to find out which is better - implement a hash map and see for yourself. You should be able to get something near constant lookup :)
I have an big array of length N, let's say something like:
2 4 6 7 6 3 3 3 4 3 4 4 4 3 3 1
I need to split this array into P subarrays (in this example, P=4 would be reasonable), such that the sum of the elements in each subarray is as close as possible to sigma, being:
sigma=(sum of all elements in original array)/P
In this example, sigma=15.
For the sake of clarity, one possible result would be:
2 4 6 7 6 3 3 3 4 3 4 4 4 3 3 1
(sums: 12,19,14,15)
I have written a very naive algorithm based in how I would do the divisions by hand, but I don't know how to impose the condition that a division whose sums are (14,14,14,14,19) is worse than one that is (15,14,16,14,16).
Thank you in advance.
First, let’s formalize your optimization problem by specifying the input, output, and the measure for each possible solution (I hope this is in your interest):
Given an array A of positive integers and a positive integer P, separate the array A into P non-overlapping subarrays such that the difference between the sum of each subarray and the perfect sum of the subarrays (sum(A)/P) is minimal.
Input: Array A of positive integers; P is a positive integer.
Output: Array SA of P non-negative integers representing the length of each subarray of A where the sum of these subarray lengths is equal to the length of A.
Measure: abs(sum(sa)-sum(A)/P) is minimal for each sa ∈ {sa | sa = (Ai, …, Ai+SAj) for i = (Σ SAj), j from 0 to P-1}.
The input and output define the set of valid solutions. The measure defines a measure to compare multiple valid solutions. And since we’re looking for a solution with the least difference to the perfect solution (minimization problem), measure should also be minimal.
With this information, it is quite easy to implement the measure function (here in Python):
def measure(a, sa):
sigma = sum(a)/len(sa)
diff = 0
i = 0
for j in xrange(0, len(sa)):
diff += abs(sum(a[i:i+sa[j]])-sigma)
i += sa[j]
return diff
print measure([2,4,6,7,6,3,3,3,4,3,4,4,4,3,3,1], [3,4,4,5]) # prints 8
Now finding an optimal solution is a little harder.
We can use the Backtracking algorithm for finding valid solutions and use the measure function to rate them. We basically try all possible combinations of P non-negative integer numbers that sum up to length(A) to represent all possible valid solutions. Although this ensures not to miss a valid solution, it is basically a brute-force approach with the benefit that we can omit some branches that cannot be any better than our yet best solution. E.g. in the example above, we wouldn’t need to test solutions with [9,…] (measure > 38) if we already have a solution with measure ≤ 38.
Following the pseudocode pattern from Wikipedia, our bt function looks as follows:
def bt(c):
global P, optimum, optimum_diff
if reject(P,c):
return
if accept(P,c):
print "%r with %d" % (c, measure(P,c))
if measure(P,c) < optimum_diff:
optimum = c
optimum_diff = measure(P,c)
return
s = first(P,c)
while s is not None:
bt(list(s))
s = next(P,s)
The global variables P, optimum, and optimum_diff represent the problem instance holding the values for A, P, and sigma, as well as the optimal solution and its measure:
class MinimalSumOfSubArraySumsProblem:
def __init__(self, a, p):
self.a = a
self.p = p
self.sigma = sum(a)/p
Next we specify the reject and accept functions that are quite straight forward:
def reject(P,c):
return optimum_diff < measure(P,c)
def accept(P,c):
return None not in c
This simply rejects any candidate whose measure is already more than our yet optimal solution. And we’re accepting any valid solution.
The measure function is also slightly changed due to the fact that c can now contain None values:
def measure(P, c):
diff = 0
i = 0
for j in xrange(0, P.p):
if c[j] is None:
break;
diff += abs(sum(P.a[i:i+c[j]])-P.sigma)
i += c[j]
return diff
The remaining two function first and next are a little more complicated:
def first(P,c):
t = 0
is_complete = True
for i in xrange(0, len(c)):
if c[i] is None:
if i+1 < len(c):
c[i] = 0
else:
c[i] = len(P.a) - t
is_complete = False
break;
else:
t += c[i]
if is_complete:
return None
return c
def next(P,s):
t = 0
for i in xrange(0, len(s)):
t += s[i]
if i+1 >= len(s) or s[i+1] is None:
if t+1 > len(P.a):
return None
else:
s[i] += 1
return s
Basically, first either replaces the next None value in the list with either 0 if it’s not the last value in the list or with the remainder to represent a valid solution (little optimization here) if it’s the last value in the list, or it return None if there is no None value in the list. next simply increments the rightmost integer by one or returns None if an increment would breach the total limit.
Now all you need is to create a problem instance, initialize the global variables and call bt with the root:
P = MinimalSumOfSubArraySumsProblem([2,4,6,7,6,3,3,3,4,3,4,4,4,3,3,1], 4)
optimum = None
optimum_diff = float("inf")
bt([None]*P.p)
If I am not mistaken here, one more approach is dynamic programming.
You can define P[ pos, n ] as the smallest possible "penalty" accumulated up to position pos if n subarrays were created. Obviously there is some position pos' such that
P[pos', n-1] + penalty(pos', pos) = P[pos, n]
You can just minimize over pos' = 1..pos.
The naive implementation will run in O(N^2 * M), where N - size of the original array and M - number of divisions.
#Gumbo 's answer is clear and actionable, but consumes lots of time when length(A) bigger than 400 and P bigger than 8. This is because that algorithm is kind of brute-forcing with benefits as he said.
In fact, a very fast solution is using dynamic programming.
Given an array A of positive integers and a positive integer P, separate the array A into P non-overlapping subarrays such that the difference between the sum of each subarray and the perfect sum of the subarrays (sum(A)/P) is minimal.
Measure: , where is sum of elements of subarray , is the average of P subarray' sums.
This can make sure the balance of sum, because it use the definition of Standard Deviation.
Persuming that array A has N elements; Q(i,j) means the minimum Measure value when split the last i elements of A into j subarrays. D(i,j) means (sum(B)-sum(A)/P)^2 when array B consists of the i~jth elements of A ( 0<=i<=j<N ).
The minimum measure of the question is to calculate Q(N,P). And we find that:
Q(N,P)=MIN{Q(N-1,P-1)+D(0,0); Q(N-2,P-1)+D(0,1); ...; Q(N-1,P-1)+D(0,N-P)}
So it like can be solved by dynamic programming.
Q(i,1) = D(N-i,N-1)
Q(i,j) = MIN{ Q(i-1,j-1)+D(N-i,N-i);
Q(i-2,j-1)+D(N-i,N-i+1);
...;
Q(j-1,j-1)+D(N-i,N-j)}
So the algorithm step is:
1. Cal j=1:
Q(1,1), Q(2,1)... Q(3,1)
2. Cal j=2:
Q(2,2) = MIN{Q(1,1)+D(N-2,N-2)};
Q(3,2) = MIN{Q(2,1)+D(N-3,N-3); Q(1,1)+D(N-3,N-2)}
Q(4,2) = MIN{Q(3,1)+D(N-4,N-4); Q(2,1)+D(N-4,N-3); Q(1,1)+D(N-4,N-2)}
... Cal j=...
P. Cal j=P:
Q(P,P), Q(P+1,P)...Q(N,P)
The final minimum Measure value is stored as Q(N,P)!
To trace each subarray's length, you can store the
MIN choice when calculate Q(i,j)=MIN{Q+D...}
space for D(i,j);
time for calculate Q(N,P)
compared to the pure brute-forcing algorithm consumes time.
Working code below (I used php language). This code decides part quantity itself;
$main = array(2,4,6,1,6,3,2,3,4,3,4,1,4,7,3,1,2,1,3,4,1,7,2,4,1,2,3,1,1,1,1,4,5,7,8,9,8,0);
$pa=0;
for($i=0;$i < count($main); $i++){
$p[]= $main[$i];
if(abs(15 - array_sum($p)) < abs(15 - (array_sum($p)+$main[$i+1])))
{
$pa=$pa+1;
$pi[] = $i+1;
$pc = count($pi);
$ba = $pi[$pc-2] ;
$part[$pa] = array_slice( $main, $ba, count($p));
unset($p);
}
}
print_r($part);
for($s=1;$s<count($part);$s++){
echo '<br>';
echo array_sum($part[$s]);
}
code will output part sums like as below
13
14
16
14
15
15
17
I'm wondering whether the following would work:
Go from the left, as soon as sum > sigma, branch into two, one including the value that pushes it over, and one that doesn't. Recursively process data to the right with rightSum = totalSum-leftSum and rightP = P-1.
So, at the start, sum = 60
2 4 6 7 6 3 3 3 4 3 4 4 4 3 3 1
Then for 2 4 6 7, sum = 19 > sigma, so split into:
2 4 6 7 6 3 3 3 4 3 4 4 4 3 3 1
2 4 6 7 6 3 3 3 4 3 4 4 4 3 3 1
Then we process 7 6 3 3 3 4 3 4 4 4 3 3 1 and 6 3 3 3 4 3 4 4 4 3 3 1 with P = 4-1 and sum = 60-12 and sum = 60-19 respectively.
This results in, I think, O(P*n).
It might be a problem when 1 or 2 values is by far the largest, but, for any value >= sigma, we can probably just put that in it's own partition (preprocessing the array to find these might be the best idea (and reduce sum appropriately)).
If it works, it should hopefully minimise sum-of-squared-error (or close to that), which seems like the desired measure.
I propose an algorithm based on backtracking. The main function chosen randomly select an element from the original array and adds it to an array partitioned. For each addition will check to obtain a better solution than the original. This will be achieved by using a function that calculates the deviation, distinguishing each adding a new element to the page. Anyway, I thought it would be good to add an original variables in loops that you can not reach desired solution will force the program ends. By desired solution I means to add all elements with respect of condition imposed by condition from if.
sum=CalculateSum(vector)
Read P
sigma=sum/P
initialize P vectors, with names vector_partition[i], i=1..P
list_vector initialize a list what pointed this P vectors
initialize a diferences_vector with dimension of P
//that can easy visualize like a vector of vectors
//construct a non-recursive backtracking algorithm
function Deviation(vector) //function for calculate deviation of elements from a vector
{
dev=0
for i=0 to Size(vector)-1 do
dev+=|vector[i+1]-vector[i]|
return dev
}
iteration=0
//fix some maximum number of iteration for while loop
Read max_iteration
//as the number of iterations will be higher the more it will get
//a more accurate solution
while(!IsEmpty(vector))
{
for i=1 to Size(list_vector) do
{
if(IsEmpty(vector)) break from while loop
initial_deviation=Deviation(list_vector[i])
el=SelectElement(vector) //you can implement that function using a randomized
//choice of element
difference_vector[i]=|sigma-CalculateSum(list_vector[i])|
PutOnBackVector(vector_list[i], el)
if(initial_deviation>Deviation(difference_vector))
ExtractFromBackVectorAndPutOnSecondVector(list_vector, vector)
}
iteration++
//prevent to enter in some infinite loop
if (iteration>max_iteration) break from while loop
}
You can change this by adding in first if some code witch increment with a amount the calculated deviation.
aditional_amount=0
iteration=0
while
{
...
if(initial_deviation>Deviation(difference_vector)+additional_amount)
ExtractFromBackVectorAndPutOnSecondVector(list_vector, vector)
if(iteration>max_iteration)
{
iteration=0
aditional_amout+=1/some_constant
}
iteration++
//delete second if from first version
}
Your problem is very similar to, or the same as, the minimum makespan scheduling problem, depending on how you define your objective. In the case that you want to minimize the maximum |sum_i - sigma|, it is exactly that problem.
As referenced in the Wikipedia article, this problem is NP-complete for p > 2. Graham's list scheduling algorithm is optimal for p <= 3, and provides an approximation ratio of 2 - 1/p. You can check out the Wikipedia article for other algorithms and their approximation.
All the algorithms given on this page are either solving for a different objective, incorrect/suboptimal, or can be used to solve any problem in NP :)
This is very similar to the case of the one-dimensional bin packing problem, see http://www.cs.sunysb.edu/~algorith/files/bin-packing.shtml. In the associated book, The Algorithm Design Manual, Skienna suggests a first-fit decreasing approach. I.e. figure out your bin size (mean = sum / N), and then allocate the largest remaining object into the first bin that has room for it. You either get to a point where you have to start over-filling a bin, or if you're lucky you get a perfect fit. As Skiena states "First-fit decreasing has an intuitive appeal to it, for we pack the bulky objects first and hope that little objects can fill up the cracks."
As a previous poster said, the problem looks like it's NP-complete, so you're not going to solve it perfectly in reasonable time, and you need to look for heuristics.
I recently needed this and did as follows;
create an initial sub-arrays array of length given sub arrays count. sub arrays should have a sum property too. ie [[sum:0],[sum:0]...[sum:0]]
sort the main array descending.
search for the sub-array with the smallest sum and insert one item from main array and increment the sub arrays sum property by the inserted item's value.
repeat item 3 up until the end of main array is reached.
return the initial array.
This is the code in JS.
function groupTasks(tasks,groupCount){
var sum = tasks.reduce((p,c) => p+c),
initial = [...Array(groupCount)].map(sa => (sa = [], sa.sum = 0, sa));
return tasks.sort((a,b) => b-a)
.reduce((groups,task) => { var group = groups.reduce((p,c) => p.sum < c.sum ? p : c);
group.push(task);
group.sum += task;
return groups;
},initial);
}
var tasks = [...Array(50)].map(_ => ~~(Math.random()*10)+1), // create an array of 100 random elements among 1 to 10
result = groupTasks(tasks,7); // distribute them into 10 sub arrays with closest sums
console.log("input array:", JSON.stringify(tasks));
console.log(result.map(r=> [JSON.stringify(r),"sum: " + r.sum]));
You can use Max Flow algorithm.
An rpc server is given which receives millions of requests a day. Each request i takes processing time Ti to get processed. We want to find the 65th percentile processing time (when processing times are sorted according to their values in increasing order) at any moment. We cannot store processing times of all the requests of the past as the number of requests is very large. And so the answer need not be exact 65th percentile, you can give some approximate answer i.e. processing time which will be around the exact 65th percentile number.
Hint: Its something to do how a histogram (i.e. an overview) is stored for a very large data without storing all of data.
Take one day's data. Use it to figure out what size to make your buckets (say one day's data shows that the vast majority (95%?) of your data is within 0.5 seconds of 1 second (ridiculous values, but hang in)
To get 65th percentile, you'll want at least 20 buckets in that range, but be generous, and make it 80. So you divide your 1 second window (-0.5 seconds to +0.5 seconds) into 80 buckets by making each 1/80th of a second wide.
Each bucket is 1/80th of 1 second. Make bucket 0 be (center - deviation) = (1 - 0.5) = 0.5 to itself + 1/80th of a second. Bucket 1 is 0.5+1/80th - 0.5 + 2/80ths. Etc.
For every value, find out which bucket it falls in, and increment a counter for that bucket.
To find 65th percentile, get the total count, and walk the buckets from zero until you get to 65% of that total.
Whenever you want to reset, set the counters all to zero.
If you always want to have good data available, keep two of these, and alternate resetting them, using the one you reset least recently as having more useful data.
Use an updown filter:
if q < x:
q += .01 * (x - q) # up a little
else:
q += .005 * (x - q) # down a little
Here a quantile estimator q tracks the x stream,
moving a little towards each x.
If both factors were .01, it would move up as often as down,
tracking the 50 th percentile.
With .01 up, .005 down, it floats up, 67 th percentile;
in general, it tracks the up / (up + down) th percentile.
Bigger up/down factors track faster but noisier --
you'll have to experiment on your real data.
(I have no idea how to analyze updowns, would appreciate a link.)
The updown() below works on long vectors X, Q in order to plot them:
#!/usr/bin/env python
from __future__ import division
import sys
import numpy as np
import pylab as pl
def updown( X, Q, up=.01, down=.01 ):
""" updown filter: running ~ up / (up + down) th percentile
here vecs X in, Q out to plot
"""
q = X[0]
for j, x in np.ndenumerate(X):
if q < x:
q += up * (x - q) # up a little
else:
q += down * (x - q) # down a little
Q[j] = q
return q
#...............................................................................
if __name__ == "__main__":
N = 1000
up = .01
down = .005
plot = 0
seed = 1
exec "\n".join( sys.argv[1:] ) # python this.py N= up= down=
np.random.seed(seed)
np.set_printoptions( 2, threshold=100, suppress=True ) # .2f
title = "updown random.exponential: N %d up %.2g down %.2g" % (N, up, down)
print title
X = np.random.exponential( size=N )
Q = np.zeros(N)
updown( X, Q, up=up, down=down )
# M = np.zeros(N)
# updown( X, M, up=up, down=up )
print "last 10 Q:", Q[-10:]
if plot:
fig = pl.figure( figsize=(8,3) )
pl.title(title)
x = np.arange(N)
pl.plot( x, X, "," )
pl.plot( x, Q )
pl.ylim( 0, 2 )
png = "updown.png"
print >>sys.stderr, "writing", png
pl.savefig( png )
pl.show()
An easier way to get the value that represents a given percentile of a list or array is the scoreatpercentile function in the scipy.stats module.
>>>import scipy.stats as ss
>>>ss.scoreatpercentile(v,65)
there's a sibling percentileofscore to return the percentile given the value
you will need to store a running sum and a total count.
then check out standard deviation calculations.