I've been trying to wrap my head around the fact that a lot of DP questions that involve bottom up tabulation via a 2D Matrix can be simplified into a 1D Array to save on space since you only rely on the previous two rows but I don't really understand the why/how/intuition behind this.
Just wondering if anyone could offer the most dumb downed version of why this works...
Generally, we use can apply DP when the optimal solution to a given problem can be determined from the optimal solutions of its subproblems. When coming up with a solution to some algorithm, it usually helps to come up with a recursive one first. From there, if we observe that recursive subproblems are re-calculated multiple times, we can just memoize the intermediate results for fast reference to them later.
In some special cases, we don't actually need to remember the solution to all subproblems at once; we just need to know a certain subset at a time.
The space optimization described above seems to best answer to the question you're asking - how does one condense the total set of solutions as a 2D matrix into a single 1D array? Well, at a given time, we don't actually store all solutions (the 2D matrix) in any single point in time; we just store what is needed to calculate the next round of intermediate/final outputs in the algorithm.
Perhaps walking through an example application may help reinforce this description.
A nice example is the generalized stock trading problem. Basically, we have an input array prices of a stock on a given day and would like to calculate the maximum profit that can be earned if k buy-sell transactions are made, where one stock may be bought and held at any given time.
The trickiest part in my opinion is figuring out how to move from the one transaction case to the two transaction case. I'll assume we're proficient enough in dynamic programming to move immediately to the k transaction case. Notice that a nice formulation of the problem in terms of subproblems is the following:
prices = input array of prices, length is n
Define dp[k][i] = maximum profit earned by day i, after having made k transactions
dp[k][i] = max(dp[k][i - 1], (prices[i] + effectivePrice)
effectivePrice = max(dp[k - 1][i] - prices[i], effectivePrice) (compute on the fly for each i)
Now in this particular case our "naive" dp solution has a 2D matrix with k rows and n columns. The space reduction here is that, in order to calculate the result for k transactions, we only need knowledge of the case for k - 1 transactions. Therefore, it is certainly possible to solve the problem using two 1D arrays of size n.
Let oldDp = solution for the k - 1 case
Let newDp = solution for the k case (computed on the fly)
for each transaction:
for each day i:
newDp[i] = max(newDp[i - 1], (prices[i] + effectivePrice)
effectivePrice = max(oldDp[i] - prices[i], effectivePrice)
// Set up for next iteration
oldDp = newDp
newDp = blank array of size n
As we can see, we managed to save a lot of space - we went from having to use a 2D matrix with k rows and n columns to two 1D arrays of size n. An even better optimization is to just use a single 1D array; this is possible since the only indices that we examine in oldDp is the current one i when calculating effectivePrice. Because we only need to temporarily remember the old result for day i, we can just make use of a temporary variable. Thus, the optimized pseudocode (for our "naive" approach) appears as below:
Let dp = maximum profit, so that dp[i] = maximum profit after k transactions (built iteratively) on day i.
for each transaction:
for each day i:
// At this point, dp[i] is equivalent to dp[k - 1][i], yet
// for all j < i, dp[j] is equivalent to dp[k][j]!
temp = dp[i]
dp[i] = max(dp[i - 1], prices[i] + effectivePrice)
effectivePrice = max(temp - prices[i], effectivePrice)
And so, using the "naive" idea of determining the optimal solution after k transactions from k - 1 transactions, we optimize space by going from a 2D matrix of size kn to a 1D array of size n.
The problem I'm having can be reduced to:
Given an array of N positive numbers, find the non-contiguous sequence of exactly K elements with the minimal sum.
Ok-ish: report the sum only. Bonus: the picked elements can be identified (at least one set of indices, if many can realize the same sum).
(in layman terms: pick any K non-neighbouring elements from N values so that their sum is minimal)
Of course, 2*K <= N+1 (otherwise no solution is possible), the problem is insensitive to positive/negative (just shift the array values with the MIN=min(A...) then add back K*MIN to the answer).
What I got so far (the naive approach):
select K+2 indexes of the values closest to the minimum. I'm not sure about this, for K=2 this seems to be the required to cover all the particular cases, but I don't know if it is required/sufficient for K>2**
brute force the minimal sum from the values of indices resulted at prev step respecting the non-contiguity criterion - if I'm right and K+2 is enough, I can live brute-forcing a (K+1)*(K+2) solution space but, as I said. I'm not sure K+2 is enough for K>2 (if in fact 2*K points are necessary, then brute-forcing goes out of window - the binomial coefficient C(2*K, K) grows prohibitively fast)
Any clever idea of how this can be done with minimal time/space complexity?
** for K=2, a non-trivial example where 4 values closest to the absolute minimum are necessary to select the objective sum [4,1,0,1,4,3,4] - one cannot use the 0 value for building the minimal sum, as it breaks the non-contiguity criterion.
PS - if you feel like showing code snippets, C/C++ and/or Java will be appreciated, but any language with decent syntax or pseudo-code will do (I reckon "decent syntax" excludes Perl, doesn't it?)
Let's assume input numbers are stored in array a[N]
Generic approach is DP: f(n, k) = min(f(n-1, k), f(n-2, k-1)+a[n])
It takes O(N*K) time and has 2 options:
for lazy backtracking recursive solution O(N*K) space
for O(K) space for forward cycle
In special case of big K there is another possibility:
use recursive back-tracking
instead of helper array of N*K size use map(n, map(k, pair(answer, list(answer indexes))))
save answer and list of indexes for this answer
instantly return MAX_INT if k>N/2
This way you'll have lower time than O(NK) for K~=N/2, something like O(Nlog(N)). It will increase up to O(N*log(N)Klog(K)) for small K, so decision between general approach or special case algorithm is important.
There should be a dynamic programming approach to this.
Work along the array from left to right. At each point i, for each value of j from 1..k, find the value of the right answer for picking j non-contiguous elements from 1..i. You can work out the answers at i by looking at the answers at i-1, i-2, and the value of array[i]. The answer you want is the answer at n for an array of length n. After you have done this you should be able to work out what the elements are by back-tracking along the array to work out whether the best decision at each point involves selecting the array element at that point, and therefore whether it used array[i-1][k] or array[i-2][k-1].
I'm working on a problem that requires an array (dA[j], j=-N..N) to be calculated from the values of another array (A[i], i=-N..N) based on a conservation of momentum rule (x+y=z+j). This means that for a given index j for all the valid combinations of (x,y,z) I calculate A[x]A[y]A[z]. dA[j] is equal to the sum of these values.
I'm currently precomputing the valid indices for each dA[j] by looping x=-N...+N,y=-N...+N and calculating z=x+y-j and storing the indices if abs(z) <= N.
Is there a more efficient method of computing this?
The reason I ask is that in future I'd like to also be able to efficiently find for each dA[j] all the terms that have a specific A[i]. Essentially to be able to compute the Jacobian of dA[j] with respect to dA[i].
Update
For the sake of completeness I figured out a way of doing this without any if statements: if you parametrize the equation x+y=z+j given that j is a constant you get the equation for a plane. The constraint that x,y,z need to be integers between -N..N create boundaries on this plane. The points that define this boundary are functions of N and j. So all you have to do is loop over your parametrized variables (s,t) within these boundaries and you'll generate all the valid points by using the vectors defined by the plane (s*u + t*v + j*[0,0,1]).
For example, if you choose u=[1,0,-1] and v=[0,1,1] all the valid solutions for every value of j are bounded by a 6 sided polygon with points (-N,-N),(-N,-j),(j,N),(N,N),(N,-j), and (j,-N).
So for each j, you go through all (2N)^2 combinations to find the correct x's and y's such that x+y= z+j; the running time of your application (per j) is O(N^2). I don't think your current idea is bad (and after playing with some pseudocode for this, I couldn't improve it significantly). I would like to note that once you've picked a j and a z, there is at most 2N choices for x's and y's. So overall, the best algorithm would still complete in O(N^2).
But consider the following improvement by a factor of 2 (for the overall program, not per j): if z+j= x+y, then (-z)+(-j)= (-x)+(-y) also.
I'm re-reading Skiena's Algorithm Design Manual to catch up on some stuff I've forgotten since school, and I'm a little baffled by his descriptions of Dynamic Programming. I've looked it up on Wikipedia and various other sites, and while the descriptions all make sense, I'm having trouble figuring out specific problems myself. Currently, I'm working on problem 3-5 from the Skiena book. (Given an array of n real numbers, find the maximum sum in any contiguous subvector of the input.) I have an O(n^2) solution, such as described in this answer. But I'm stuck on the O(N) solution using dynamic programming. It's not clear to me what the recurrence relation should be.
I see that the subsequences form a set of sums, like so:
S = {a,b,c,d}
a a+b a+b+c a+b+c+d
b b+c b+c+d
c c+d
d
What I don't get is how to pick which one is the greatest in linear time. I've tried doing things like keeping track of the greatest sum so far, and if the current value is positive, add it to the sum. But when you have larger sequences, this becomes problematic because there may be stretches of negative numbers that would decrease the sum, but a later large positive number may bring it back to being the maximum.
I'm also reminded of summed area tables. You can calculate all the sums using only the cumulative sums: a, a+b, a+b+c, a+b+c+d, etc. (For example, if you need b+c, it's just (a+b+c) - (a).) But don't see an O(N) way to get it.
Can anyone explain to me what the O(N) dynamic programming solution is for this particular problem? I feel like I almost get it, but that I'm missing something.
You should take a look to this pdf back in the school in http://castle.eiu.edu here it is:
The explanation of the following pseudocode is also int the pdf.
There is a solution like, first sort the array in to some auxiliary memory, then apply Longest Common Sub-Sequence method to the original array and the sorted array, with sum(not the length) of common sub-sequence in the 2 arrays as the entry into the table (Memoization). This can also solve the problem
Total running time is O(nlogn)+O(n^2) => O(n^2)
Space is O(n) + O(n^2) => O(n^2)
This is not a good solution when memory comes into picture. This is just to give a glimpse on how problems can be reduced to one another.
My understanding of DP is about "making a table". In fact, the original meaning "programming" in DP is simply about making tables.
The key is to figure out what to put in the table, or modern terms: what state to track, or what's the vertex key/value in DAG (ignore these terms if they sound strange to you).
How about choose dp[i] table being the largest sum ending at index i of the array, for example, the array being [5, 15, -30, 10]
The second important key is "optimal substructure", that is to "assume" dp[i-1] already stores the largest sum for sub-sequences ending at index i-1, that's why the only step at i is to decide whether to include a[i] into the sub-sequence or not
dp[i] = max(dp[i-1], dp[i-1] + a[i])
The first term in max is to "not include a[i]", the second term is to "include a[i]". Notice, if we don't include a[i], the largest sum so far remains dp[i-1], which comes from the "optimal substructure" argument.
So the whole program looks like this (in Python):
a = [5,15,-30,10]
dp = [0]*len(a)
dp[0] = max(0,a[0]) # include a[0] or not
for i in range(1,len(a)):
dp[i] = max(dp[i-1], dp[i-1]+a[i]) # for sub-sequence, choose to add or not
print(dp, max(dp))
The result: largest sum of sub-sequence should be the largest item in dp table, after i iterate through the array a. But take a close look at dp, it holds all the information.
Since it only goes through items in array a once, it's a O(n) algorithm.
This problem seems silly, because as long as a[i] is positive, we should always include it in the sub-sequence, because it will only increase the sum. This intuition matches the code
dp[i] = max(dp[i-1], dp[i-1] + a[i])
So the max. sum of sub-sequence problem is easy, and doesn't need DP at all. Simply,
sum = 0
for v in a:
if v >0
sum += v
However, what about largest sum of "continuous sub-array" problem. All we need to change is just a single line of code
dp[i] = max(dp[i-1]+a[i], a[i])
The first term is to "include a[i] in the continuous sub-array", the second term is to decide to start a new sub-array, starting a[i].
In this case, dp[i] is the max. sum continuous sub-array ending with index-i.
This is certainly better than a naive approach O(n^2)*O(n), to for j in range(0,i): inside the i-loop and sum all the possible sub-arrays.
One small caveat, because the way dp[0] is set, if all items in a are negative, we won't select any. So for the max sum continuous sub-array, we change that to
dp[0] = a[0]
Given an n×n matrix of real numbers. You are allowed to erase any number (from 0 to n) of rows and any number (from 0 to n) of columns, and after that the sum of the remaining entries is computed. Come up with an algorithm which finds out which rows and columns to erase in order to maximize that sum.
The problem is NP-hard. (So you should not expect a polynomial-time algorithm for solving this problem. There could still be (non-polynomial time) algorithms that are slightly better than brute-force, though.) The idea behind the proof of NP-hardness is that if we could solve this problem, then we could solve the the clique problem in a general graph. (The maximum-clique problem is to find the largest set of pairwise connected vertices in a graph.)
Specifically, given any graph with n vertices, let's form the matrix A with entries a[i][j] as follows:
a[i][j] = 1 for i == j (the diagonal entries)
a[i][j] = 0 if the edge (i,j) is present in the graph (and i≠j)
a[i][j] = -n-1 if the edge (i,j) is not present in the graph.
Now suppose we solve the problem of removing some rows and columns (or equivalently, keeping some rows and columns) so that the sum of the entries in the matrix is maximized. Then the answer gives the maximum clique in the graph:
Claim: In any optimal solution, there is no row i and column j kept for which the edge (i,j) is not present in the graph. Proof: Since a[i][j] = -n-1 and the sum of all the positive entries is at most n, picking (i,j) would lead to a negative sum. (Note that deleting all rows and columns would give a better sum, of 0.)
Claim: In (some) optimal solution, the set of rows and columns kept is the same. This is because starting with any optimal solution, we can simply remove all rows i for which column i has not been kept, and vice-versa. Note that since the only positive entries are the diagonal ones, we do not decrease the sum (and by the previous claim, we do not increase it either).
All of which means that if the graph has a maximum clique of size k, then our matrix problem has a solution with sum k, and vice-versa. Therefore, if we could solve our initial problem in polynomial time, then the clique problem would also be solved in polynomial time. This proves that the initial problem is NP-hard. (Actually, it is easy to see that the decision version of the initial problem — is there a way of removing some rows and columns so that the sum is at least k — is in NP, so the (decision version of the) initial problem is actually NP-complete.)
Well the brute force method goes something like this:
For n rows there are 2n subsets.
For n columns there are 2n subsets.
For an n x n matrix there are 22n subsets.
0 elements is a valid subset but obviously if you have 0 rows or 0 columns the total is 0 so there are really 22n-2+1 subsets but that's no different.
So you can work out each combination by brute force as an O(an) algorithm. Fast. :)
It would be quicker to work out what the maximum possible value is and you do that by adding up all the positive numbers in the grid. If those numbers happen to form a valid sub-matrix (meaning you can create that set by removing rows and/or columns) then there's your answer.
Implicit in this is that if none of the numbers are negative then the complete matrix is, by definition, the answer.
Also, knowing what the highest possible maximum is possibly allows you to shortcut the brute force evaluation since if you get any combination equal to that maximum then that is your answer and you can stop checking.
Also if all the numbers are non-positive, the answer is the maximum value as you can reduce the matrix to a 1 x 1 matrix with that 1 value in it, by definition.
Here's an idea: construct 2n-1 n x m matrices where 1 <= m <= n. Process them one after the other. For each n x m matrix you can calculate:
The highest possible maximum sum (as per above); and
Whether no numbers are positive allowing you to shortcut the answer.
if (1) is below the currently calculate highest maximum sum then you can discard this n x m matrix. If (2) is true then you just need a simple comparison to the current highest maximum sum.
This is generally referred to as a pruning technique.
What's more you can start by saying that the highest number in the n x n matrix is the starting highest maximum sum since obviously it can be a 1 x 1 matrix.
I'm sure you could tweak this into a (slightly more) efficient recursive tree-based search algorithm with the above tests effectively allowing you to eliminate (hopefully many) unnecessary searches.
We can improve on Cletus's generalized brute-force solution by modelling this as a directed graph. The initial matrix is the start node of the graph; its leaves are all the matrices missing one row or column, and so forth. It's a graph rather than a tree, because the node for the matrix without both the first column and row will have two parents - the nodes with just the first column or row missing.
We can optimize our solution by turning the graph into a tree: There's never any point exploring a submatrix with a column or row deleted that comes before the one we deleted to get to the current node, as that submatrix will be arrived at anyway.
This is still a brute-force search, of course - but we've eliminated the duplicate cases where we remove the same rows in different orders.
Here's an example implementation in Python:
def maximize_sum(m):
frontier = [(m, 0, False)]
best = None
best_score = 0
while frontier:
current, startidx, cols_done = frontier.pop()
score = matrix_sum(current)
if score > best_score or not best:
best = current
best_score = score
w, h = matrix_size(current)
if not cols_done:
for x in range(startidx, w):
frontier.append((delete_column(current, x), x, False))
startidx = 0
for y in range(startidx, h):
frontier.append((delete_row(current, y), y, True))
return best_score, best
And here's the output on 280Z28's example matrix:
>>> m = ((1, 1, 3), (1, -89, 101), (1, 102, -99))
>>> maximize_sum(m)
(106, [(1, 3), (1, 101)])
Since nobody asked for an efficient algorithm, use brute force: generate every possible matrix that can be created by removing rows and/or columns from the original matrix, choose the best one. A slightly more efficent version, which most likely can be proved to still be correct, is to generate only those variants where the removed rows and columns contain at least one negative value.
To try it in a simple way:
We need the valid subset of the set of entries {A00, A01, A02, ..., A0n, A10, ...,Ann} which max. sum.
First compute all subsets (the power set).
A valid subset is a member of the power set that for each two contained entries Aij and A(i+x)(j+y), contains also the elements A(i+x)j and Ai(j+y) (which are the remaining corners of the rectangle spanned by Aij and A(i+x)(j+y)).
Aij ...
. .
. .
... A(i+x)(j+y)
By that you can eliminate the invalid ones from the power set and find the one with the biggest sum in the remaining.
I'm sure it can be improved by improving an algorithm for power set generation in order to generate only valid subsets and by that avoiding step 2 (adjusting the power set).
I think there are some angles of attack that might improve upon brute force.
memoization, since there are many distinct sequences of edits that will arrive at the same submatrix.
dynamic programming. Because the search space of matrices is highly redundant, my intuition is that there would be a DP formulation that can save a lot of repeated work
I think there's a heuristic approach, but I can't quite nail it down:
if there's one negative number, you can either take the matrix as it is, remove the column of the negative number, or remove its row; I don't think any other "moves" result in a higher sum. For two negative numbers, your options are: remove neither, remove one, remove the other, or remove both (where the act of removal is either by axing the row or the column).
Now suppose the matrix has only one positive number and the rest are all <=0. You clearly want to remove everything but the positive entry. For a matrix with only 2 positive entries and the rest <= 0, the options are: do nothing, whittle down to one, whittle down to the other, or whittle down to both (resulting in a 1x2, 2x1, or 2x2 matrix).
In general this last option falls apart (imagine a matrix with 50 positives & 50 negatives), but depending on your data (few negatives or few positives) it could provide a shortcut.
Create an n-by-1 vector RowSums, and an n-by-1 vector ColumnSums. Initialize them to the row and column sums of the original matrix. O(n²)
If any row or column has a negative sum, remove edit: the one with the minimum such and update the sums in the other direction to reflect their new values. O(n)
Stop when no row or column has a sum less than zero.
This is an iterative variation improving on another answer. It operates in O(n²) time, but fails for some cases mentioned in other answers, which is the complexity limit for this problem (there are n² entries in the matrix, and to even find the minimum you have to examine each cell once).
Edit: The following matrix has no negative rows or columns, but is also not maximized, and my algorithm doesn't catch it.
1 1 3 goal 1 3
1 -89 101 ===> 1 101
1 102 -99
The following matrix does have negative rows and columns, but my algorithm selects the wrong ones for removal.
-5 1 -5 goal 1
1 1 1 ===> 1
-10 2 -10 2
mine
===> 1 1 1
Compute the sum of each row and column. This can be done in O(m) (where m = n^2)
While there are rows or columns that sum to negative remove the row or column that has the lowest sum that is less than zero. Then recompute the sum of each row/column.
The general idea is that as long as there is a row or a column that sums to nevative, removing it will result in a greater overall value. You need to remove them one at a time and recompute because in removing that one row/column you are affecting the sums of the other rows/columns and they may or may not have negative sums any more.
This will produce an optimally maximum result. Runtime is O(mn) or O(n^3)
I cannot really produce an algorithm on top of my head, but to me it 'smells' like dynamic programming, if it serves as a start point.
Big Edit: I honestly don't think there's a way to assess a matrix and determine it is maximized, unless it is completely positive.
Maybe it needs to branch, and fathom all elimination paths. You never no when a costly elimination will enable a number of better eliminations later. We can short circuit if it's found the theoretical maximum, but other than any algorithm would have to be able to step forward and back. I've adapted my original solution to achieve this behaviour with recursion.
Double Secret Edit: It would also make great strides to reduce to complexity if each iteration didn't need to find all negative elements. Considering that they don't change much between calls, it makes more sense to just pass their positions to the next iteration.
Takes a matrix, the list of current negative elements in the matrix, and the theoretical maximum of the initial matrix. Returns the matrix's maximum sum and the list of moves required to get there. In my mind move list contains a list of moves denoting the row/column removed from the result of the previous operation.
Ie: r1,r1
Would translate
-1 1 0 1 1 1
-4 1 -4 5 7 1
1 2 4 ===>
5 7 1
Return if sum of matrix is the theoretical maximum
Find the positions of all negative elements unless an empty set was passed in.
Compute sum of matrix and store it along side an empty move list.
For negative each element:
Calculate the sum of that element's row and column.
clone the matrix and eliminate which ever collection has the minimum sum (row/column) from that clone, note that action as a move list.
clone the list of negative elements and remove any that are effected by the action taken in the previous step.
Recursively call this algorithm providing the cloned matrix, the updated negative element list and the theoretical maximum. Append the moves list returned to the move list for the action that produced the matrix passed to the recursive call.
If the returned value of the recursive call is greater than the stored sum, replace it and store the returned move list.
Return the stored sum and move list.
I'm not sure if it's better or worse than the brute force method, but it handles all the test cases now. Even those where the maximum contains negative values.
This is an optimization problem and can be solved approximately by an iterative algorithm based on simulated annealing:
Notation: C is number of columns.
For J iterations:
Look at each column and compute the absolute benefit of toggling it (turn it off if it's currently on or turn it on if it's currently off). That gives you C values, e.g. -3, 1, 4. A greedy deterministic solution would just pick the last action (toggle the last column to get a benefit of 4) because it locally improves the objective. But that might lock us into a local optimum. Instead, we probabilistically pick one of the three actions, with probabilities proportional to the benefits. To do this, transform them into a probability distribution by putting them through a Sigmoid function and normalizing. (Or use exp() instead of sigmoid()?) So for -3, 1, 4 you get 0.05, 0.73, 0.98 from the sigmoid and 0.03, 0.42, 0.56 after normalizing. Now pick the action according to the probability distribution, e.g. toggle the last column with probability 0.56, toggle the second column with probability 0.42, or toggle the first column with the tiny probability 0.03.
Do the same procedure for the rows, resulting in toggling one of the rows.
Iterate for J iterations until convergence.
We may also, in early iterations, make each of these probability distributions more uniform, so that we don't get locked into bad decisions early on. So we'd raise the unnormalized probabilities to a power 1/T, where T is high in early iterations and is slowly decreased until it approaches 0. For example, 0.05, 0.73, 0.98 from above, raised to 1/10 results in 0.74, 0.97, 1.0, which after normalization is 0.27, 0.36, 0.37 (so it's much more uniform than the original 0.05, 0.73, 0.98).
It's clearly NP-Complete (as outlined above). Given this, if I had to propose the best algorithm I could for the problem:
Try some iterations of quadratic integer programming, formulating the problem as: SUM_ij a_ij x_i y_j, with the x_i and y_j variables constrained to be either 0 or 1. For some matrices I think this will find a solution quickly, for the hardest cases it would be no better than brute force (and not much would be).
In parallel (and using most of the CPU), use a approximate search algorithm to generate increasingly better solutions. Simulating Annealing was suggested in another answer, but having done research on similar combinatorial optimisation problems, my experience is that tabu search would find good solutions faster. This is probably close to optimal in terms of wandering between distinct "potentially better" solutions in the shortest time, if you use the trick of incrementally updating the costs of single changes (see my paper "Graph domination, tabu search and the football pool problem").
Use the best solution so far from the second above to steer the first by avoiding searching possibilities that have lower bounds worse than it.
Obviously this isn't guaranteed to find the maximal solution. But, it generally would when this is feasible, and it would provide a very good locally maximal solution otherwise. If someone had a practical situation requiring such optimisation, this is the solution that I'd think would work best.
Stopping at identifying that a problem is likely to be NP-Complete will not look good in a job interview! (Unless the job is in complexity theory, but even then I wouldn't.) You need to suggest good approaches - that is the point of a question like this. To see what you can come up with under pressure, because the real world often requires tackling such things.
yes, it's NP-complete problem.
It's hard to easily find the best sub-matrix,but we can easily to find some better sub-matrix.
Assume that we give m random points in the matrix as "feeds". then let them to automatically extend by the rules like :
if add one new row or column to the feed-matrix, ensure that the sum will be incrementive.
,then we can compare m sub-matrix to find the best one.
Let's say n = 10.
Brute force (all possible sets of rows x all possible sets of columns) takes
2^10 * 2^10 =~ 1,000,000 nodes.
My first approach was to consider this a tree search, and use
the sum of positive entries is an upper bound for every node in the subtree
as a pruning method. Combined with a greedy algorithm to cheaply generate good initial bounds, this yielded answers in about 80,000 nodes on average.
but there is a better way ! i later realised that
Fix some choice of rows X.
Working out the optimal columns for this set of rows is now trivial (keep a column if its sum of its entries in the rows X is positive, otherwise discard it).
So we can just brute force over all possible choices of rows; this takes 2^10 = 1024 nodes.
Adding the pruning method brought this down to 600 nodes on average.
Keeping 'column-sums' and incrementally updating them when traversing the tree of row-sets should allow the calculations (sum of matrix etc) at each node to be O(n) instead of O(n^2). Giving a total complexity of O(n * 2^n)
For slightly less than optimal solution, I think this is a PTIME, PSPACE complexity issue.
The GREEDY algorithm could run as follows:
Load the matrix into memory and compute row totals. After that run the main loop,
1) Delete the smallest row,
2) Subtract the newly omitted values from the old row totals
--> Break when there are no more negative rows.
Point two is a subtle detail: subtracted two rows/columns has time complexity n.
While re-summing all but two columns has n^2 time complexity!
Take each row and each column and compute the sum. For a 2x2 matrix this will be:
2 1
3 -10
Row(0) = 3
Row(1) = -7
Col(0) = 5
Col(1) = -9
Compose a new matrix
Cost to take row Cost to take column
3 5
-7 -9
Take out whatever you need to, then start again.
You just look for negative values on the new matrix. Those are values that actually substract from the overall matrix value. It terminates when there're no more negative "SUMS" values to take out (therefore all columns and rows SUM something to the final result)
In an nxn matrix that would be O(n^2)Log(n) I think
function pruneMatrix(matrix) {
max = -inf;
bestRowBitField = null;
bestColBitField = null;
for(rowBitField=0; rowBitField<2^matrix.height; rowBitField++) {
for (colBitField=0; colBitField<2^matrix.width; colBitField++) {
sum = calcSum(matrix, rowBitField, colBitField);
if (sum > max) {
max = sum;
bestRowBitField = rowBitField;
bestColBitField = colBitField;
}
}
}
return removeFieldsFromMatrix(bestRowBitField, bestColBitField);
}
function calcSumForCombination(matrix, rowBitField, colBitField) {
sum = 0;
for(i=0; i<matrix.height; i++) {
for(j=0; j<matrix.width; j++) {
if (rowBitField & 1<<i && colBitField & 1<<j) {
sum += matrix[i][j];
}
}
}
return sum;
}