This problem appeared in code jam 2018 qualification round which has ended.
https://codejam.withgoogle.com/2018/challenges/ (Problem 2)
Problem description:
The basic operation of the standard bubble sort algorithm is to examine a pair of adjacent numbers and reverse that pair if the left number is larger than the right number. But our algorithm examines a group of three adjacent numbers, and if the leftmost number is larger than the rightmost number, it reverses that entire group. Because our algorithm is a "triplet bubble sort", we have named it Trouble Sort for short.
We were looking forward to presenting Trouble Sort at the Special
Interest Group in Sorting conference in Hawaii, but one of our interns
has just pointed out a problem: it is possible that Trouble Sort does
not correctly sort the list! Consider the list 8 9 7, for example.
We need your help with some further research. Given a list of N
integers, determine whether Trouble Sort will successfully sort the
list into non-decreasing order. If it will not, find the index
(counting starting from 0) of the first sorting error after the
algorithm has finished: that is, the first value that is larger than
the value that comes directly after it when the algorithm is done.
So a naive approach will be to apply trouble sort on the given list, apply normal sort on the list, and find the index of the first non-matching element. However, this would time out for very large N.
Here is what I figured:
The algorithm will compare 0th index with 2nd, 2nd with 4th and so on.
Similarly 1st with 3rd, 3rd with 5th and so on.
All the elements at odd index will be sorted with respect to odd index. Same for even indexed element.
So the issue would lie between two consecutive odd/even indexed element.
I can't think of a way to figure it out without doing an O(n^2) approach.
Is my approach any viable, or there is something easier?
Your observation is spot on. The algorithm presented in the problem statement will only compare( and swap ) the consecutive odd and even elements among themselves.
If you take that observation one step further, you can state that Trouble Sort is an algorithm that correctly sorts odd- and even-indexed elements of an array within themselves. (i.e. as if odd-indexed elements and even-indexed elements of an array A are two separate arrays B and C)
In other words, Trouble Sort does sort B and C correctly. The issue here is whether those arrays B and C of odd and even-indexed elements can be merged properly. You should check if sorting odd- and even-indexed elements among themselves is enough to make the entire array sorted.
This step is really similar to the merging step of MergeSort. The only difference is that, due to the indexing being a limiting factor on your operation, you know at all times from which array you will pick the top element. For a 1-indexed array A, during the merging step of B and C, at each step, you should pick the smallest previously unpicked element from B, and then C.
So, basically, if you sort B and C, which takes, O(NlogN) using an algorithm such as mergesort or heapsort, and then merge them in the manner described in the previous paragraph, which takes O(N), you end up with the same version of the array A after it has been processed by the Trouble Sort algorithm.
The difference is the time complexity. While Trouble Sort takes O(N^2) time, the operations described above takes O(NlogN) time. Once you end up with this array, then you can check in O(N) time if, for each consecutive indices i, j, A[i] < A[j] holds. The overall complexity of the algorithm would still be O(NlogN).
Below is a code sample in Python to demonstrate sort of a pseudocode of the algorithm I described above. There are a couple of minor differences in implementation due to Python arrays being 0-indexed. You may observe the execution of this code here.
def does_trouble_sort_work(A):
B, C = A[0::2], A[1::2]
B_sorted = sorted(B)
C_sorted = sorted(C)
j = k = 0
for i in xrange(len(A)):
if i % 2 == 0:
A[i] = B_sorted[j]
j += 1
else:
A[i] = C_sorted[k]
k += 1
trouble_sort_works = True
for i in xrange(1, len(A)):
if A[i-1] > A[i]:
trouble_sort_works = False
break
return trouble_sort_works
I have n integers a_1, ..., a_n. I want to pick the minimum number from all of them whose xor forms others.
For example, consider [1,2,3], 1^3=2 so you don't need 2 in the array. So you can remove it. To end up with [1,3]. So the min number of elements is 2 and they can form all the original elements in the array by xoring any 2 of them. Would a greedy approach work here? or DP?
Edit: To explain what I am thinking. A greedy approach I thought about was due to the fact that if a^b=c then a^c=b and b^c=a. First I delete all duplicates. then I would first in the beginning list all the pairs that each element can pair up with to form another element in the array. It takes O(n^3) for preprocessing. Then I pick the element with the least contribution and I delete it and subsequently subtract 1 from each of the other elements. I repeat this until all elements have <=2 pairs. and I stop. This would also take O(n^3) for a total of O(n^3). Does this greedy approach work? Is there a DP way to do it?
If n is bounded by 50 I think backtracking should work.
Suppose at some step we have already selected a subset S of numbers (that should produce all the others) and want to include a new number to that subset.
Then we can do the following:
Consider all remaining numbers R and include in S all numbers that can't be produced by others (in S and R)
Include in S a random (or "best" in some way) number from R
Remove from R all numbers that can be produced by those in updated S
Also you should keep track of the current best solution and cut off all the branches that won't allow to get a better result.
In a recent campus Facebook interview i have asked to divide an array into 3 equal parts such that the sum in each array is roughly equal to sum/3.My Approach1. Sort The Array2. Fill the array[k] (k=0) uptil (array[k]<=sum/3)3. After that increment k and repeat the above step for array[k]Is there any better algorithm for this or it is NP Hard Problem
This is a variant of the partition problem (see http://en.wikipedia.org/wiki/Partition_problem for details). In fact a solution to this can solve that one (take an array, pad with 0s, and then solve this problem) so this problem is NP hard.
There is a dynamic programming approach that is pseudo-polynomial. For each i from 0 to the size of the array, you keep track of all possible combinations of current sizes for the sub arrays, and their current sums. As long as there are a limited number possible sums of subsets of the array, this runs acceptably fast.
The solution that I would have suggested is to just go for "good enough" closeness. First let's consider the simpler problem with all values positive. Then sort by value descending. Take that array in threes. Build up the three subsets by always adding the largest of the triple to the one with the smallest sum, the smallest to the one with the largest, and the middle to the middle. You will end up dividing the array evenly, and the difference will be no more than the value of the third smallest element.
For the general case you can divide into positive and negative, use the above approach on each, and then brute force all combinations of a group of positives, a group of negatives, and the few leftover values in the middle that did not divide evenly.
Here are details on a dynamic programming solution if you are interested. The running time and memory usage is O(n*(sum)^2) where n is the size of your array and sum is the sum of absolute values of your array values. For each array index j from 1 to n, store all the possible values you can get for your 3 subset sums when you split the array from index 1 to j into 3 subsets. Also for each possibility, store one possible way to split the array to get the 3 sums. Then to extend this information for 1 to (j+1) given the information from 1 to j, simply take each possible combination of 3 sums for splitting 1 to j and form the 3 combinations of 3 sums you get when you choose to add the (j+1)th array element to any one of the 3 subsets. Finally, when you are done and reach j = n, go through the set of all combinations of 3 subset sums you can get when you split array positions 1 to n into 3 sets, and choose the one whose maximum deviation from sum/3 is minimized. At first this may seem like O(n*(sum)^3) complexity, but for each j and each combination of the first 2 subset sums, the 3rd subset sum is uniquely determined. (because you are not allowed to omit any elements of the array). Thus the complexity really is O(n*(sum)^2).
Given is a array of numbers:
1, 2, 8, 6, 9, 0, 4
We need to find all the numbers in group of three which sums to a value N ( say 11 in this example). Here, the possible numbers in group of three are:
{1,2,8}, {1,4,6}, {0,2,9}
The first solution I could think was of O(n^3). Later I could improve a little(n^2 log n) with the approach:
1. Sort the array.
2. Select any two number and perform binary search for the third element.
Can it be improved further with some other approaches?
You can certainly do it in O(n^2): for each i in the array, test whether two other values sum to N-i.
You can test in O(n) whether two values in a sorted array sum to k by sweeping from both ends at once. If the sum of the two elements you're on is too big, decrement the "right-to-left" index to make it smaller. If the sum is too small, increment the "left-to-right" index to make it bigger. If there's a pair that works, you'll find them, and you perform at most 2*n iterations before you run out of road at one end or the other. You might need code to ignore the value you're using as i, depends what the rules are.
You could instead use some kind of dynamic programming, working down from N, and you probably end up with time something like O(n*N) or so. Realistically I don't think that's any better: it looks like all your numbers are non-negative, so if n is much bigger than N then before you start you can quickly throw out any large values from the array, and also any duplicates beyond 3 copies of each value (or 2 copies, as long as you check whether 3*i == N before discarding the 3rd copy of i). After that step, n is O(N).
Is there a way to generate all of the subset sums s1, s2, ..., sk that fall in a range [A,B] faster than O((k+N)*2N/2), where k is the number of sums there are in [A,B]? Note that k is only known after we have enumerated all subset sums within [A,B].
I'm currently using a modified Horowitz-Sahni algorithm. For example, I first call it to for the smallest sum greater than or equal to A, giving me s1. Then I call it again for the next smallest sum greater than s1, giving me s2. Repeat this until we find a sum sk+1 greater than B. There is a lot of computation repeated between each iteration, even without rebuilding the initial two 2N/2 lists, so is there a way to do better?
In my problem, N is about 15, and the magnitude of the numbers is on the order of millions, so I haven't considered the dynamic programming route.
Check the subset sum on Wikipedia. As far as I know, it's the fastest known algorithm, which operates in O(2^(N/2)) time.
Edit:
If you're looking for multiple possible sums, instead of just 0, you can save the end arrays and just iterate through them again (which is roughly an O(2^(n/2) operation) and save re-computing them. The value of all the possible subsets is doesn't change with the target.
Edit again:
I'm not wholly sure what you want. Are we running K searches for one independent value each, or looking for any subset that has a value in a specific range that is K wide? Or are you trying to approximate the second by using the first?
Edit in response:
Yes, you do get a lot of duplicate work even without rebuilding the list. But if you don't rebuild the list, that's not O(k * N * 2^(N/2)). Building the list is O(N * 2^(N/2)).
If you know A and B right now, you could begin iteration, and then simply not stop when you find the right answer (the bottom bound), but keep going until it goes out of range. That should be roughly the same as solving subset sum for just one solution, involving only +k more ops, and when you're done, you can ditch the list.
More edit:
You have a range of sums, from A to B. First, you solve subset sum problem for A. Then, you just keep iterating and storing the results, until you find the solution for B, at which point you stop. Now you have every sum between A and B in a single run, and it will only cost you one subset sum problem solve plus K operations for K values in the range A to B, which is linear and nice and fast.
s = *i + *j; if s > B then ++i; else if s < A then ++j; else { print s; ... what_goes_here? ... }
No, no, no. I get the source of your confusion now (I misread something), but it's still not as complex as what you had originally. If you want to find ALL combinations within the range, instead of one, you will just have to iterate over all combinations of both lists, which isn't too bad.
Excuse my use of auto. C++0x compiler.
std::vector<int> sums;
std::vector<int> firstlist;
std::vector<int> secondlist;
// Fill in first/secondlist.
std::sort(firstlist.begin(), firstlist.end());
std::sort(secondlist.begin(), secondlist.end());
auto firstit = firstlist.begin();
auto secondit = secondlist.begin();
// Since we want all in a range, rather than just the first, we need to check all combinations. Horowitz/Sahni is only designed to find one.
for(; firstit != firstlist.end(); firstit++) {
for(; secondit = secondlist.end(); secondit++) {
int sum = *firstit + *secondit;
if (sum > A && sum < B)
sums.push_back(sum);
}
}
It's still not great. But it could be optimized if you know in advance that N is very large, for example, mapping or hashmapping sums to iterators, so that any given firstit can find any suitable partners in secondit, reducing the running time.
It is possible to do this in O(N*2^(N/2)), using ideas similar to Horowitz Sahni, but we try and do some optimizations to reduce the constants in the BigOh.
We do the following
Step 1: Split into sets of N/2, and generate all possible 2^(N/2) sets for each split. Call them S1 and S2. This we can do in O(2^(N/2)) (note: the N factor is missing here, due to an optimization we can do).
Step 2: Next sort the larger of S1 and S2 (say S1) in O(N*2^(N/2)) time (we optimize here by not sorting both).
Step 3: Find Subset sums in range [A,B] in S1 using binary search (as it is sorted).
Step 4: Next, for each sum in S2, find using binary search the sets in S1 whose union with this gives sum in range [A,B]. This is O(N*2^(N/2)). At the same time, find if that corresponding set in S2 is in the range [A,B]. The optimization here is to combine loops. Note: This gives you a representation of the sets (in terms of two indexes in S2), not the sets themselves. If you want all the sets, this becomes O(K + N*2^(N/2)), where K is the number of sets.
Further optimizations might be possible, for instance when sum from S2, is negative, we don't consider sums < A etc.
Since Steps 2,3,4 should be pretty clear, I will elaborate further on how to get Step 1 done in O(2^(N/2)) time.
For this, we use the concept of Gray Codes. Gray codes are a sequence of binary bit patterns in which each pattern differs from the previous pattern in exactly one bit.
Example: 00 -> 01 -> 11 -> 10 is a gray code with 2 bits.
There are gray codes which go through all possible N/2 bit numbers and these can be generated iteratively (see the wiki page I linked to), in O(1) time for each step (total O(2^(N/2)) steps), given the previous bit pattern, i.e. given current bit pattern, we can generate the next bit pattern in O(1) time.
This enables us to form all the subset sums, by using the previous sum and changing that by just adding or subtracting one number (corresponding to the differing bit position) to get the next sum.
If you modify the Horowitz-Sahni algorithm in the right way, then it's hardly slower than original Horowitz-Sahni. Recall that Horowitz-Sahni works two lists of subset sums: Sums of subsets in the left half of the original list, and sums of subsets in the right half. Call these two lists of sums L and R. To obtain subsets that sum to some fixed value A, you can sort R, and then look up a number in R that matches each number in L using a binary search. However, the algorithm is asymmetric only to save a constant factor in space and time. It's a good idea for this problem to sort both L and R.
In my code below I also reverse L. Then you can keep two pointers into R, updated for each entry in L: A pointer to the last entry in R that's too low, and a pointer to the first entry in R that's too high. When you advance to the next entry in L, each pointer might either move forward or stay put, but they won't have to move backwards. Thus, the second stage of the Horowitz-Sahni algorithm only takes linear time in the data generated in the first stage, plus linear time in the length of the output. Up to a constant factor, you can't do better than that (once you have committed to this meet-in-the-middle algorithm).
Here is a Python code with example input:
# Input
terms = [29371, 108810, 124019, 267363, 298330, 368607,
438140, 453243, 515250, 575143, 695146, 840979, 868052, 999760]
(A,B) = (500000,600000)
# Subset iterator stolen from Sage
def subsets(X):
yield []; pairs = []
for x in X:
pairs.append((2**len(pairs),x))
for w in xrange(2**(len(pairs)-1), 2**(len(pairs))):
yield [x for m, x in pairs if m & w]
# Modified Horowitz-Sahni with toolow and toohigh indices
L = sorted([(sum(S),S) for S in subsets(terms[:len(terms)/2])])
R = sorted([(sum(S),S) for S in subsets(terms[len(terms)/2:])])
(toolow,toohigh) = (-1,0)
for (Lsum,S) in reversed(L):
while R[toolow+1][0] < A-Lsum and toolow < len(R)-1: toolow += 1
while R[toohigh][0] <= B-Lsum and toohigh < len(R): toohigh += 1
for n in xrange(toolow+1,toohigh):
print '+'.join(map(str,S+R[n][1])),'=',sum(S+R[n][1])
"Moron" (I think he should change his user name) raises the reasonable issue of optimizing the algorithm a little further by skipping one of the sorts. Actually, because each list L and R is a list of sizes of subsets, you can do a combined generate and sort of each one in linear time! (That is, linear in the lengths of the lists.) L is the union of two lists of sums, those that include the first term, term[0], and those that don't. So actually you should just make one of these halves in sorted form, add a constant, and then do a merge of the two sorted lists. If you apply this idea recursively, you save a logarithmic factor in the time to make a sorted L, i.e., a factor of N in the original variable of the problem. This gives a good reason to sort both lists as you generate them. If you only sort one list, you have some binary searches that could reintroduce that factor of N; at best you have to optimize them somehow.
At first glance, a factor of O(N) could still be there for a different reason: If you want not just the subset sum, but the subset that makes the sum, then it looks like O(N) time and space to store each subset in L and in R. However, there is a data-sharing trick that also gets rid of that factor of O(N). The first step of the trick is to store each subset of the left or right half as a linked list of bits (1 if a term is included, 0 if it is not included). Then, when the list L is doubled in size as in the previous paragraph, the two linked lists for a subset and its partner can be shared, except at the head:
0
|
v
1 -> 1 -> 0 -> ...
Actually, this linked list trick is an artifact of the cost model and never truly helpful. Because, in order to have pointers in a RAM architecture with O(1) cost, you have to define data words with O(log(memory)) bits. But if you have data words of this size, you might as well store each word as a single bit vector rather than with this pointer structure. I.e., if you need less than a gigaword of memory, then you can store each subset in a 32-bit word. If you need more than a gigaword, then you have a 64-bit architecture or an emulation of it (or maybe 48 bits), and you can still store each subset in one word. If you patch the RAM cost model to take account of word size, then this factor of N was never really there anyway.
So, interestingly, the time complexity for the original Horowitz-Sahni algorithm isn't O(N*2^(N/2)), it's O(2^(N/2)). Likewise the time complexity for this problem is O(K+2^(N/2)), where K is the length of the output.