Consider the following problem:
I have sequence with 2q elements {a[1], ..., a[2q]}, which has been sorted. Then I need to get two sub-sequences b[i] and c[i], which satisfy:
Both b[i] and c[i] have q elements and Join({b[i]}, {c[i]}) = {a[i]}
b[1] < b[2] < ... < b[q]
b[i] < c[i] for all i = 1...q
The difficulty is that I want to get all sequence that satisfy the conditions. How can I do that?
The first step is to partition the input into two halves, while making sure that each partition will lead to a solution. We can do that with a simple recursive method:
Iterate over the input from small to large; for each element, choose whether it goes into subsequence b or c, using these rules:
If b and c have the same length, add the element to b.
if b is longer than c, add the element once to b and once to c and recurse with both options.
If b is full, add all elements to c.
This will give us a number if different partitions, e.g. for the input [1,2,3,4,5,6] they would be:
[1,2,3],[4,5,6]
[1,2,4],[3,5,6]
[1,2,5],[3,4,6]
[1,3,4],[2,5,6]
[1,3,5],[2,4,6]
Then, for each of these partitions, we have to find the permutations of c which meet the conditions. Again, a simple recursive method will do the trick:
Iterate over b from right to left. For each element, find the elements in c which are greater. If there is only one, put it in the corresponding location in c. If there are severeal, put each of the elements in the corresponding location in c once, and recurse with each option.
For the partition [1,2,4],[3,5,6] in the example, that would lead to:
[1,2,4],[x,x,x] <- (3,5,6) ... [1,2,4],[x,x,6] <- (3,5) ... [1,2,4],[3,5,6]
^ ^ ^ ^ ^ ^
... [1,2,4],[x,x,6] <- (3,5) ... [1,2,4],[5,3,6]
^ ^ ^
[1,2,4],[x,x,x] <- (3,5,6) ... [1,2,4],[x,x,5] <- (3,6) ... [1,2,4],[3,6,5]
^ ^ ^ ^ ^ ^
... [1,2,4],[x,x,5] <- (3,6) ... [1,2,4],[6,3,5]
^ ^ ^
This code example is a straightforward implementation of the algorithm explained above:
function subsequences(a) {
var q = a.length / 2;
partition(0, [], []);
function partition(pos, part1, part2) {
var b = part1.slice(); // create copy
var c = part2.slice(); // create copy
if (b.length == c.length) {
b.push(a[pos++]);
}
while (b.length < q) {
c.push(a[pos++]);
partition(pos, b, c);
b.push(c.pop());
}
while (c.length < q) {
c.push(a[pos++]);
}
permute(b, [], c);
}
function permute(b, part, set) {
var pos = set.length - 1;
for (var i = pos; i >= 0 && set[i] > b[pos]; i--) {
var c = part.slice(); // create copy
var s = set.slice(); // create copy
c[pos] = s.splice(i, 1);
if (pos == 0) { // store or process subsequences
document.write("{" + b + "},{" + c + "}<br>");
}
else permute(b, c, s);
}
}
}
subsequences([1,2,3,4,5,6,7,8,9,10,11,12]); // ascending order
The algorithm finds the following number of valid partitions, and valid permutations of c. (For comparison, I've added the number of all partitions and all permutations that a naive solution would need to generate and then check.)
q partitions permutations | all part. all perm.
|
1 1 1 | 1 1
2 2 3 | 3 6
3 5 15 | 10 60
4 14 105 | 35 840
5 42 945 | 126 15,120
6 132 10,395 | 462 332,640
7 429 135,135 | 1,716 8,648,640
8 1,430 2,027,025 | 6,435 259,459,200
9 4,862 34,459,425 | 24,310 8,821,612,800
10 16,796 654,729,075 | 92,378 335,221,286,400
Running only the first part of the algorithm, we find that for q=20, the number of valid partitions is 6,564,120,420, and the number of valid permutations is probably around 1022.
NOTES:
For the partition function to work correctly, the input sequence needs to be in ascending order. In the permute function, the set of unused values in subsequence c is also kept in ascending order, to make it easier to find the largest values.
The results for any input sequence of n numbers will be similar. In fact, you can replace the input sequence by [0,1,2,3 ... n-1] and then use the values in the results as indexes in the original input sequence. This means that if you need the results for different sequences of n numbers, you only need to run the algorithm once.
Related
I'm solving the exercise 2.1-4 from CLRS "introduction to algorithms".
The problem is described as:
Consider the problem of adding two n-bit binary integers, stored in two n-element arrays A and B. The sum of the two integers should be stored in binary form in element array C.
What is the loop invariant of this problem?
I have some thoughts about this question, and wrote them as comments in my solution to this problem, written in golang.
package additoin_binary
/*
Loop invariant:
At the start of each iteration of the loop digits in the subarray r[len(r) - 1 - i:] are
a[len(a) - 1 - i:] + b[len(b) - 1 - i:] + carry | provided that (len(a) - 1 - i) and (len(b) - 1 - i) are positive
a[len(a) - 1 - i:] + carry | provided that (len(a) - 1 - i) is positive and (len(b) - 1 - i) is negative
carry | provided that (len(a) - 1 - i) and (len(b) - 1 - i) are negative
*** carry for i = a[(len(a) - 1) - i - 1] + b[(len(b) - 1) - i - 1] == 2 ? 1 : 0
*/
func BinaryAddition(a, b []int) []int {
// a and b can have arbitrary number of digits.
// We should control a length of the second term. It should be less or equal to a length of first term.w
// Other vise we swap them
if len(b) > len(a) {
b, a = a, b
}
// loop invariant initialization:
// before first loop iteration (i=0) index b_i is out of the array range (b[-1]), so we don't have second term and sum = a
// another way of thinking about it is we don't get b as argument and then sum = a, too
if len(b) == 0 {
return a
}
// result array to store sum
r := make([]int, len(a)+1)
// overflow of summing two bits (1 + 1)
carry := 0
// loop invariant maintenance:
// we have right digits (after addition) in r for indexes r[len(r) - 1 - i:]
for i := 0; i < len(r); i++ {
a_i := len(a) - 1 - i // index for getting a last digit of a
b_i := len(b) - 1 - i // index for getting a last digit of b
r_i := len(r) - 1 - i // index for getting a last digit of r
var s int
if b_i >= 0 && a_i >= 0 {
s = a[a_i] + b[b_i] + carry
} else if a_i >= 0 {
s = a[a_i] + carry
} else { // all indexes run out of the game (a < 0, b < 0)
s = carry
}
if s > 1 {
r[r_i] = 0
carry = 1
} else {
r[r_i] = s
carry = 0
}
}
// loop invariant termination:
// i goes from 0 to len(r) - 1, r[len(r) - 1 - ([len(r) - 1):] => r[:]
// This means, that for every index in r we have a right sum
//*At i=0, r[i] a sum can be equal to 0, so we explicitly check that before return r
if r[0] == 0 {
return r[1:]
} else {
return r
}
}
Edit 1: I extended the original problem. So now arrays A and B can have arbitrary lengths, respectively m and n. Example A = [1,0,1], B = [1,0] (m=3, n=2)
Consider the problem of adding two n-bit binary integers, stored in two n-element arrays A and B. The sum of the two integers should be stored in binary form in element array C.
The problem has a guarantee that A and B are n-element arrays, I think it's an important condition which could reduce code work.
What is a loop invariant?
In simple words, a loop invariant is some predicate (condition) that holds for every iteration of the loop.
In this problem, if assume len = len(C), iterate i in [0, len), the loop invariant is that r[len-1-i:len] is always the sum of a[len-2-i:len-1] and b[len-2-i:len-1] in lower i + 1 bit. Because after each loop, you will make one bit correct, it could prove that algorithm is correct.
The loop invariant condition can be taken as the number of bits yet to be added n - p (assuming you start by adding the lsb bits first from right to left), where p I have taken as the current bit significant position and n the size of the Augend and Addend bit-sequences.
Given the number of rows and columns of a 2d matrix
Initially all elements of matrix are 0
Given the number of 1's that should be present in each row
Given the number of 1's that should be present in each column
Determine if it is possible to form such matrix.
Example:
Input: r=3 c=2 (no. of rows and columns)
2 1 0 (number of 1's that should be present in each row respectively)
1 2 (number of 1's that should be present in each column respectively)
Output: Possible
Explanation:
1 1
0 1
0 0
I tried solving this problem for like 12 hours by checking if summation of Ri = summation of Ci
But I wondered if wouldn't be possible for cases like
3 3
1 3 0
0 2 2
r and c can be upto 10^5
Any ideas how should I move further?
Edit: Constraints added and output should only be "possible" or "impossible". The possible matrix need not be displayed.
Can anyone help me now?
Hint: one possible solution utilizes Maximum Flow Problem by creating a special graph and running the standard maximum flow algorithm on it.
If you're not familiar with the above problem, you may start reading about it e.g. here https://en.wikipedia.org/wiki/Maximum_flow_problem
If you're interested in the full solution please comment and I'll update the answer. But it requires understading the above algorithm.
Solution as requested:
Create a graph of r+c+2 nodes.
Node 0 is the source, node r+c+1 is the sink. Nodes 1..r represent the rows, while r+1..r+c the columns.
Create following edges:
from source to nodes i=1..r of capacity r_i
from nodes i=r+1..r+c to sink of capacity c_i
between all the nodes i=1..r and j=r+1..r+c of capacity 1
Run maximum flow algorithm, the saturated edges between row nodes and column nodes define where you should put 1.
Or if it's not possible then the maximum flow value is less than number of expected ones in the matrix.
I will illustrate the algorithm with an example.
Assume we have m rows and n columns. Let rows[i] be the number of 1s in row i, for 0 <= i < m,
and cols[j] be the number of 1s in column j, for 0 <= j < n.
For example, for m = 3, and n = 4, we could have: rows = {4 2 3}, cols = {1 3 2 3}, and
the solution array would be:
1 3 2 3
+--------
4 | 1 1 1 1
2 | 0 1 0 1
3 | 0 1 1 1
Because we only want to know whether a solution exists, the values in rows and cols may be permuted in any order. The solution of each permutation is just a permutation of the rows and columns of the above solution.
So, given rows and cols, sort cols in decreasing order, and rows in increasing order. For our example, we have cols = {3 3 2 1} and rows = {2 3 4}, and the equivalent problem.
3 3 2 1
+--------
2 | 1 1 0 0
3 | 1 1 1 0
4 | 1 1 1 1
We transform cols into a form that is better suited for the algorithm. What cols tells us is that we have two series of 1s of length 3, one series of 1s of length 2, and one series of 1s of length 1, that are to be distributed among the rows of the array. We rewrite cols to capture just that, that is COLS = {2/3 1/2 1/1}, 2 series of length 3, 1 series of length 2, and 1 series of length 1.
Because we have 2 series of length 3, a solution exists only if we can put two 1s in the first row. This is possible because rows[0] = 2. We do not actually put any 1 in the first row, but record the fact that 1s have been placed there by decrementing the length of the series of length 3. So COLS becomes:
COLS = {2/2 1/2 1/1}
and we combine our two counts for series of length 2, yielding:
COLS = {3/2 1/1}
We now have the reduced problem:
3 | 1 1 1 0
4 | 1 1 1 1
Again we need to place 1s from our series of length 2 to have a solution. Fortunately, rows[1] = 3 and we can do this. We decrement the length of 3/2 and get:
COLS = {3/1 1/1} = {4/1}
We have the reduced problem:
4 | 1 1 1 1
Which is solved by 4 series of length 1, just what we have left. If at any step, the series in COLS cannot be used to satisfy a row count, then no solution is possible.
The general processing for each row may be stated as follows. For each row r, starting from the first element in COLS, decrement the lengths of as many elements count[k]/length[k] of COLS as needed, so that the sum of the count[k]'s equals rows[r]. Eliminate series of length 0 in COLS and combine series of same length.
Note that because elements of COLS are in decreasing order of lengths, the length of the last element decremented is always less than or equal to the next element in COLS (if there is a next element).
EXAMPLE 2 : Solution exists.
rows = {1 3 3}, cols = {2 2 2 1} => COLS = {3/2 1/1}
1 series of length 2 is decremented to satisfy rows[0] = 1, and the 2 other series of length 2 remains at length 2.
rows[0] = 1
COLS = {2/2 1/1 1/1} = {2/2 2/1}
The 2 series of length 2 are decremented, and 1 of the series of length 1.
The series whose length has become 0 is deleted, and the series of length 1 are combined.
rows[1] = 3
COLS = {2/1 1/0 1/1} = {2/1 1/1} = {3/1}
A solution exists for rows[2] can be satisfied.
rows[2] = 3
COLS = {3/0} = {}
EXAMPLE 3: Solution does not exists.
rows = {0 2 3}, cols = {3 2 0 0} => COLS = {1/3 1/2}
rows[0] = 0
COLS = {1/3 1/2}
rows[1] = 2
COLS = {1/2 1/1}
rows[2] = 3 => impossible to satisfy; no solution.
SPACE COMPLEXITY
It is easy to see that it is O(m + n).
TIME COMPLEXITY
We iterate over each row only once. For each row i, we need to iterate over at most
rows[i] <= n elements of COLS. Time complexity is O(m x n).
After finding this algorithm, I found the following theorem:
The Havel-Hakimi theorem (Havel 1955, Hakimi 1962) states that there exists a matrix Xn,m of 0’s and 1’s with row totals a0=(a1, a2,… , an) and column totals b0=(b1, b2,… , bm) such that bi ≥ bi+1 for every 0 < i < m if and only if another matrix Xn−1,m of 0’s and 1’s with row totals a1=(a2, a3,… , an) and column totals b1=(b1−1, b2−1,… ,ba1−1, ba1+1,… , bm) also exists.
from the post Finding if binary matrix exists given the row and column sums.
This is basically what my algorithm does, while trying to optimize the decrementing part, i.e., all the -1's in the above theorem. Now that I see the above theorem, I know my algorithm is correct. Nevertheless, I checked the correctness of my algorithm by comparing it with a brute-force algorithm for arrays of up to 50 cells.
Here is the C# implementation.
public class Pair
{
public int Count;
public int Length;
}
public class PairsList
{
public LinkedList<Pair> Pairs;
public int TotalCount;
}
class Program
{
static void Main(string[] args)
{
int[] rows = new int[] { 0, 0, 1, 1, 2, 2 };
int[] cols = new int[] { 2, 2, 0 };
bool success = Solve(cols, rows);
}
static bool Solve(int[] cols, int[] rows)
{
PairsList pairs = new PairsList() { Pairs = new LinkedList<Pair>(), TotalCount = 0 };
FillAllPairs(pairs, cols);
for (int r = 0; r < rows.Length; r++)
{
if (rows[r] > 0)
{
if (pairs.TotalCount < rows[r])
return false;
if (pairs.Pairs.First != null && pairs.Pairs.First.Value.Length > rows.Length - r)
return false;
DecrementPairs(pairs, rows[r]);
}
}
return pairs.Pairs.Count == 0 || pairs.Pairs.Count == 1 && pairs.Pairs.First.Value.Length == 0;
}
static void DecrementPairs(PairsList pairs, int count)
{
LinkedListNode<Pair> pair = pairs.Pairs.First;
while (count > 0 && pair != null)
{
LinkedListNode<Pair> next = pair.Next;
if (pair.Value.Count == count)
{
pair.Value.Length--;
if (pair.Value.Length == 0)
{
pairs.Pairs.Remove(pair);
pairs.TotalCount -= count;
}
else if (pair.Next != null && pair.Next.Value.Length == pair.Value.Length)
{
pair.Value.Count += pair.Next.Value.Count;
pairs.Pairs.Remove(pair.Next);
next = pair;
}
count = 0;
}
else if (pair.Value.Count < count)
{
count -= pair.Value.Count;
pair.Value.Length--;
if (pair.Value.Length == 0)
{
pairs.Pairs.Remove(pair);
pairs.TotalCount -= pair.Value.Count;
}
else if(pair.Next != null && pair.Next.Value.Length == pair.Value.Length)
{
pair.Value.Count += pair.Next.Value.Count;
pairs.Pairs.Remove(pair.Next);
next = pair;
}
}
else // pair.Value.Count > count
{
Pair p = new Pair() { Count = count, Length = pair.Value.Length - 1 };
pair.Value.Count -= count;
if (p.Length > 0)
{
if (pair.Next != null && pair.Next.Value.Length == p.Length)
pair.Next.Value.Count += p.Count;
else
pairs.Pairs.AddAfter(pair, p);
}
else
pairs.TotalCount -= count;
count = 0;
}
pair = next;
}
}
static int FillAllPairs(PairsList pairs, int[] cols)
{
List<Pair> newPairs = new List<Pair>();
int c = 0;
while (c < cols.Length && cols[c] > 0)
{
int k = c++;
if (cols[k] > 0)
pairs.TotalCount++;
while (c < cols.Length && cols[c] == cols[k])
{
if (cols[k] > 0) pairs.TotalCount++;
c++;
}
newPairs.Add(new Pair() { Count = c - k, Length = cols[k] });
}
LinkedListNode<Pair> pair = pairs.Pairs.First;
foreach (Pair p in newPairs)
{
while (pair != null && p.Length < pair.Value.Length)
pair = pair.Next;
if (pair == null)
{
pairs.Pairs.AddLast(p);
}
else if (p.Length == pair.Value.Length)
{
pair.Value.Count += p.Count;
pair = pair.Next;
}
else // p.Length > pair.Value.Length
{
pairs.Pairs.AddBefore(pair, p);
}
}
return c;
}
}
(Note: to avoid confusion between when I'm talking about the actual numbers in the problem vs. when I'm talking about the zeros in the ones in the matrix, I'm going to instead fill the matrix with spaces and X's. This obviously doesn't change the problem.)
Some observations:
If you're filling in a row, and there's (for example) one column needing 10 more X's and another column needing 5 more X's, then you're sometimes better off putting the X in the "10" column and saving the "5" column for later (because you might later run into 5 rows that each need 2 X's), but you're never better off putting the X in the "5" column and saving the "10" column for later (because even if you later run into 10 rows that all need an X, they won't mind if they don't all go in the same column). So we can use a somewhat "greedy" algorithm: always put an X in the column still needing the most X's. (Of course, we'll need to make sure that we don't greedily put an X in the same column multiple times for the same row!)
Since you don't need to actually output a possible matrix, the rows are all interchangeable and the columns are all interchangeable; all that matter is how many rows still need 1 X, how many still need 2 X's, etc., and likewise for columns.
With that in mind, here's one fairly simple approach:
(Optimization.) Add up the counts for all the rows, add up the counts for all the columns, and return "impossible" if the sums don't match.
Create an array of length r+1 and populate it with how many columns need 1 X, how many need 2 X's, etc. (You can ignore any columns needing 0 X's.)
(Optimization.) To help access the array efficiently, build a stack/linked-list/etc. of the indices of nonzero array elements, in decreasing order (e.g., starting at index r if it's nonzero, then index r−1 if it's nonzero, etc.), so that you can easily find the elements representing columns to put X's in.
(Optimization.) To help determine when there'll be a row can't be satisfied, also make note of the total number of columns needing any X's, and make note of the largest number of X's needed by any row. If the former is less than the latter, return "impossible".
(Optimization.) Sort the rows by the number of X's they need.
Iterate over the rows, starting with the one needing the fewest X's and ending with the one needing the most X's, and for each one:
Update the array accordingly. For example, if a row needs 12 X's, and the array looks like [..., 3, 8, 5], then you'll update the array to look like [..., 3+7 = 10, 8+5−7 = 6, 5−5 = 0]. If it's not possible to update the array because you run out of columns to put X's in, return "impossible". (Note: this part should never actually return "impossible", because we're keeping count of the number of columns left and the max number of columns we'll need, so we should have already returned "impossible" if this was going to happen. I mention this check only for clarity.)
Update the stack/linked-list of indices of nonzero array elements.
Update the total number of columns needing any X's. If it's now less than the greatest number of X's needed by any row, return "impossible".
(Optimization.) If the first nonzero array element has an index greater than the number of rows left, return "impossible".
If we complete our iteration without having returned "impossible", return "possible".
(Note: the reason I say to start with the row needing the fewest X's, and work your way to the row with the most X's, is that a row needing more X's may involve examining updating more elements of the array and of the stack, so the rows needing fewer X's are cheaper. This isn't just a matter of postponing the work: the rows needing fewer X's can help "consolidate" the array, so that there will be fewer distinct column-counts, making the later rows cheaper than they would otherwise be. In a very-bad-case scenario, such as the case of a square matrix where every single row needs a distinct positive number of X's and every single column needs a distinct positive number of X's, the fewest-to-most order means you can handle each row in O(1) time, for linear time overall, whereas the most-to-fewest order would mean that each row would take time proportional to the number of X's it needs, for quadratic time overall.)
Overall, this takes no worse than O(r+c+n) time (where n is the number of X's); I think that the optimizations I've listed are enough to ensure that it's closer to O(r+c) time, but it's hard to be 100% sure. I recommend trying it to see if it's fast enough for your purposes.
You can use brute force (iterating through all 2^(r * c) possibilities) to solve it, but that will take a long time. If r * c is under 64, you can accelerate it to a certain extent using bit-wise operations on 64-bit integers; however, even then, iterating through all 64-bit possibilities would take, at 1 try per ms, over 500M years.
A wiser choice is to add bits one by one, and only continue placing bits if no constraints are broken. This will eliminate the vast majority of possibilities, greatly speeding up the process. Look up backtracking for the general idea. It is not unlike solving sudokus through guesswork: once it becomes obvious that your guess was wrong, you erase it and try guessing a different digit.
As with sudokus, there are certain strategies that can be written into code and will result in speedups when they apply. For example, if the sum of 1s in rows is different from the sum of 1s in columns, then there are no solutions.
If over 50% of the bits will be on, you can instead work on the complementary problem (transform all ones to zeroes and vice-versa, while updating row and column counts). Both problems are equivalent, because any answer for one is also valid for the complementary.
This problem can be solved in O(n log n) using Gale-Ryser Theorem. (where n is the maximum of lengths of the two degree sequences).
First, make both sequences of equal length by adding 0's to the smaller sequence, and let this length be n.
Let the sequences be A and B. Sort A in non-decreasing order, and sort B in non-increasing order. Create another prefix sum array P for B such that ith element of P is equal to sum of first i elements of B.
Now, iterate over k's from 1 to n, and check for
The second sum can be calculated in O(log n) using binary search for index of last number in B smaller than k, and then using precalculated P.
Inspiring from the solution given by RobertBaron I have tried to build a new algorithm.
rows = [int(x)for x in input().split()]
cols = [int (ss) for ss in input().split()]
rows.sort()
cols.sort(reverse=True)
for i in range(len(rows)):
for j in range(len(cols)):
if(rows[i]!= 0 and cols[j]!=0):
rows[i] = rows[i] - 1;
cols[j] =cols[j]-1;
print("rows: ",rows)
print("cols: ",cols)
#if there is any non zero value, print NO else print yes
flag = True
for i in range(len(rows)):
if(rows[i]!=0):
flag = False
break
for j in range(len(cols)):
if(cols[j]!=0):
flag = False
if(flag):
print("YES")
else:
print("NO")
here, i have sorted the rows in ascending order and cols in descending order. later decrementing particular row and column if 1 need to be placed!
it is working for all the test cases posted here! rest GOD knows
I recently encountered a much more difficult variation of this problem, but realized I couldn't generate a solution for this very simple case. I searched Stack Overflow but couldn't find a resource that previously answered this.
You are given a triangle ABC, and you must compute the number of paths of certain length that start at and end at 'A'. Say our function f(3) is called, it must return the number of paths of length 3 that start and end at A: 2 (ABA,ACA).
I'm having trouble formulating an elegant solution. Right now, I've written a solution that generates all possible paths, but for larger lengths, the program is just too slow. I know there must be a nice dynamic programming solution that reuses sequences that we've previously computed but I can't quite figure it out. All help greatly appreciated.
My dumb code:
def paths(n,sequence):
t = ['A','B','C']
if len(sequence) < n:
for node in set(t) - set(sequence[-1]):
paths(n,sequence+node)
else:
if sequence[0] == 'A' and sequence[-1] == 'A':
print sequence
Let PA(n) be the number of paths from A back to A in exactly n steps.
Let P!A(n) be the number of paths from B (or C) to A in exactly n steps.
Then:
PA(1) = 1
PA(n) = 2 * P!A(n - 1)
P!A(1) = 0
P!A(2) = 1
P!A(n) = P!A(n - 1) + PA(n - 1)
= P!A(n - 1) + 2 * P!A(n - 2) (for n > 2) (substituting for PA(n-1))
We can solve the difference equations for P!A analytically, as we do for Fibonacci, by noting that (-1)^n and 2^n are both solutions of the difference equation, and then finding coefficients a, b such that P!A(n) = a*2^n + b*(-1)^n.
We end up with the equation P!A(n) = 2^n/6 + (-1)^n/3, and PA(n) being 2^(n-1)/3 - 2(-1)^n/3.
This gives us code:
def PA(n):
return (pow(2, n-1) + 2*pow(-1, n-1)) / 3
for n in xrange(1, 30):
print n, PA(n)
Which gives output:
1 1
2 0
3 2
4 2
5 6
6 10
7 22
8 42
9 86
10 170
11 342
12 682
13 1366
14 2730
15 5462
16 10922
17 21846
18 43690
19 87382
20 174762
21 349526
22 699050
23 1398102
24 2796202
25 5592406
26 11184810
27 22369622
28 44739242
29 89478486
The trick is not to try to generate all possible sequences. The number of them increases exponentially so the memory required would be too great.
Instead, let f(n) be the number of sequences of length n beginning and ending A, and let g(n) be the number of sequences of length n beginning with A but ending with B. To get things started, clearly f(1) = 1 and g(1) = 0. For n > 1 we have f(n) = 2g(n - 1), because the penultimate letter will be B or C and there are equal numbers of each. We also have g(n) = f(n - 1) + g(n - 1) because if a sequence ends begins A and ends B the penultimate letter is either A or C.
These rules allows you to compute the numbers really quickly using memoization.
My method is like this:
Define DP(l, end) = # of paths end at end and having length l
Then DP(l,'A') = DP(l-1, 'B') + DP(l-1,'C'), similar for DP(l,'B') and DP(l,'C')
Then for base case i.e. l = 1 I check if the end is not 'A', then I return 0, otherwise return 1, so that all bigger states only counts those starts at 'A'
Answer is simply calling DP(n, 'A') where n is the length
Below is a sample code in C++, you can call it with 3 which gives you 2 as answer; call it with 5 which gives you 6 as answer:
ABCBA, ACBCA, ABABA, ACACA, ABACA, ACABA
#include <bits/stdc++.h>
using namespace std;
int dp[500][500], n;
int DP(int l, int end){
if(l<=0) return 0;
if(l==1){
if(end != 'A') return 0;
return 1;
}
if(dp[l][end] != -1) return dp[l][end];
if(end == 'A') return dp[l][end] = DP(l-1, 'B') + DP(l-1, 'C');
else if(end == 'B') return dp[l][end] = DP(l-1, 'A') + DP(l-1, 'C');
else return dp[l][end] = DP(l-1, 'A') + DP(l-1, 'B');
}
int main() {
memset(dp,-1,sizeof(dp));
scanf("%d", &n);
printf("%d\n", DP(n, 'A'));
return 0;
}
EDITED
To answer OP's comment below:
Firstly, DP(dynamic programming) is always about state.
Remember here our state is DP(l,end), represents the # of paths having length l and ends at end. So to implement states using programming, we usually use array, so DP[500][500] is nothing special but the space to store the states DP(l,end) for all possible l and end (That's why I said if you need a bigger length, change the size of array)
But then you may ask, I understand the first dimension which is for l, 500 means l can be as large as 500, but how about the second dimension? I only need 'A', 'B', 'C', why using 500 then?
Here is another trick (of C/C++), the char type indeed can be used as an int type by default, which value is equal to its ASCII number. And I do not remember the ASCII table of course, but I know that around 300 will be enough to represent all the ASCII characters, including A(65), B(66), C(67)
So I just declare any size large enough to represent 'A','B','C' in the second dimension (that means actually 100 is more than enough, but I just do not think that much and declare 500 as they are almost the same, in terms of order)
so you asked what DP[3][1] means, it means nothing as the I do not need / calculate the second dimension when it is 1. (Or one can think that the state dp(3,1) does not have any physical meaning in our problem)
In fact, I always using 65, 66, 67.
so DP[3][65] means the # of paths of length 3 and ends at char(65) = 'A'
You can do better than the dynamic programming/recursion solution others have posted, for the given triangle and more general graphs. Whenever you are trying to compute the number of walks in a (possibly directed) graph, you can express this in terms of the entries of powers of a transfer matrix. Let M be a matrix whose entry m[i][j] is the number of paths of length 1 from vertex i to vertex j. For a triangle, the transfer matrix is
0 1 1
1 0 1.
1 1 0
Then M^n is a matrix whose i,j entry is the number of paths of length n from vertex i to vertex j. If A corresponds to vertex 1, you want the 1,1 entry of M^n.
Dynamic programming and recursion for the counts of paths of length n in terms of the paths of length n-1 are equivalent to computing M^n with n multiplications, M * M * M * ... * M, which can be fast enough. However, if you want to compute M^100, instead of doing 100 multiplies, you can use repeated squaring: Compute M, M^2, M^4, M^8, M^16, M^32, M^64, and then M^64 * M^32 * M^4. For larger exponents, the number of multiplies is about c log_2(exponent).
Instead of using that a path of length n is made up of a path of length n-1 and then a step of length 1, this uses that a path of length n is made up of a path of length k and then a path of length n-k.
We can solve this with a for loop, although Anonymous described a closed form for it.
function f(n){
var as = 0, abcs = 1;
for (n=n-3; n>0; n--){
as = abcs - as;
abcs *= 2;
}
return 2*(abcs - as);
}
Here's why:
Look at one strand of the decision tree (the other one is symmetrical):
A
B C...
A C
B C A B
A C A B B C A C
B C A B B C A C A C A B B C A B
Num A's Num ABC's (starting with first B on the left)
0 1
1 (1-0) 2
1 (2-1) 4
3 (4-1) 8
5 (8-3) 16
11 (16-5) 32
Cleary, we can't use the strands that end with the A's...
You can write a recursive brute force solution and then memoize it (aka top down dynamic programming). Recursive solutions are more intuitive and easy to come up with. Here is my version:
# search space (we have triangle with nodes)
nodes = ["A", "B", "C"]
#cache # memoize!
def recurse(length, steps):
# if length of the path is n and the last node is "A", then it's
# a valid path and we can count it.
if length == n and ((steps-1)%3 == 0 or (steps+1)%3 == 0):
return 1
# we don't want paths having len > n.
if length > n:
return 0
# from each position, we have two possibilities, either go to next
# node or previous node. Total paths will be sum of both the
# possibilities. We do this recursively.
return recurse(length+1, steps+1) + recurse(length+1, steps-1)
I wrote recursive backtracking algorithm for finding all subsets of a given set.
void backtracke(int* a, int k, int n)
{
if (k == n)
{
for(int i = 1; i <=k; ++i)
{
if (a[i] == true)
{
std::cout << i << " ";
}
}
std::cout << std::endl;
return;
}
bool c[2];
c[0] = false;
c[1] = true;
++k;
for(int i = 0; i < 2; ++i)
{
a[k] = c[i];
backtracke(a, k, n);
a[k] = INT_MAX;
}
}
now we have to write the same algorithm but in an iterative form, how to do it ?
You can use the binary counter approach. Any unique binary string of length n represents a unique subset of a set of n elements. If you start with 0 and end with 2^n-1, you cover all possible subsets. The counter can be easily implemented in an iterative manner.
The code in Java:
public static void printAllSubsets(int[] arr) {
byte[] counter = new byte[arr.length];
while (true) {
// Print combination
for (int i = 0; i < counter.length; i++) {
if (counter[i] != 0)
System.out.print(arr[i] + " ");
}
System.out.println();
// Increment counter
int i = 0;
while (i < counter.length && counter[i] == 1)
counter[i++] = 0;
if (i == counter.length)
break;
counter[i] = 1;
}
}
Note that in Java one can use BitSet, which makes the code really shorter, but I used a byte array to illustrate the process better.
There are a few ways to write an iterative algorithm for this problem. The most commonly suggested would be to:
Count (i.e. a simply for-loop) from 0 to 2numberOfElements - 1
If we look at the variable used above for counting in binary, the digit at each position could be thought of a flag indicating whether or not the element at the corresponding index in the set should be included in this subset. Simply loop over each bit (by taking the remainder by 2, then dividing by 2), including the corresponding elements in our output.
Example:
Input: {1,2,3,4,5}.
We'd start counting at 0, which is 00000 in binary, which means no flags are set, so no elements are included (this would obviously be skipped if you don't want the empty subset) - output {}.
Then 1 = 00001, indicating that only the last element would be included - output {5}.
Then 2 = 00010, indicating that only the second last element would be included - output {4}.
Then 3 = 00011, indicating that the last two elements would be included - output {4,5}.
And so on, all the way up to 31 = 11111, indicating that all the elements would be included - output {1,2,3,4,5}.
* Actually code-wise, it would be simpler to turn this on its head - output {1} for 00001, considering that the first remainder by 2 will then correspond to the flag of the 0th element, the second remainder, the 1st element, etc., but the above is simpler for illustrative purposes.
More generally, any recursive algorithm could be changed to an iterative one as follows:
Create a loop consisting of parts (think switch-statement), with each part consisting of the code between any two recursive calls in your function
Create a stack where each element contains each necessary local variable in the function, and an indication of which part we're busy with
The loop would pop elements from the stack, executing the appropriate section of code
Each recursive call would be replaced by first adding it's own state to the stack, and then the called state
Replace return with appropriate break statements
A little Python implementation of George's algorithm. Perhaps it will help someone.
def subsets(S):
l = len(S)
for x in range(2**l):
yield {s for i,s in enumerate(S) if ((x / 2**i) % 2) // 1 == 1}
Basically what you want is P(S) = S_0 U S_1 U ... U S_n where S_i is a set of all sets contained by taking i elements from S. In other words if S= {a, b, c} then S_0 = {{}}, S_1 = {{a},{b},{c}}, S_2 = {{a, b}, {a, c}, {b, c}} and S_3 = {a, b, c}.
The algorithm we have so far is
set P(set S) {
PS = {}
for i in [0..|S|]
PS = PS U Combination(S, i)
return PS
}
We know that |S_i| = nCi where |S| = n. So basically we know that we will be looping nCi times. You may use this information to optimize the algorithm later on. To generate combinations of size i the algorithm that I present is as follows:
Suppose S = {a, b, c} then you can map 0 to a, 1 to b and 2 to c. And perumtations to these are (if i=2) 0-0, 0-1, 0-2, 1-0, 1-1, 1-2, 2-0, 2-1, 2-2. To check if a sequence is a combination you check if the numbers are all unique and that if you permute the digits the sequence doesn't appear elsewhere, this will filter the above sequence to just 0-1, 0-2 and 1-2 which are later mapped back to {a,b},{a,c},{b,c}. How to generate the long sequence above you can follow this algorithm
set Combination(set S, integer l) {
CS = {}
for x in [0..2^l] {
n = {}
for i in [0..l] {
n = n U {floor(x / |S|^i) mod |S|} // get the i-th digit in x base |S|
}
CS = CS U {S[n]}
}
return filter(CS) // filtering described above
}
This is more of a puzzle than a coding problem. I need to find how many binary numbers can be generated satisfying certain constraints. The inputs are
(integer) Len - Number of digits in the binary number
(integer) x
(integer) y
The binary number has to be such that taking any x adjacent digits from the binary number should contain at least y 1's.
For example -
Len = 6, x = 3, y = 2
0 1 1 0 1 1 - Length is 6, Take any 3 adjacent digits from this and
there will be 2 l's
I had this C# coding question posed to me in an interview and I cannot figure out any algorithm to solve this. Not looking for code (although it's welcome), any sort of help, pointers are appreciated
This problem can be solved using dynamic programming. The main idea is to group the binary numbers according to the last x-1 bits and the length of each binary number. If appending a bit sequence to one number yields a number satisfying the constraint, then appending the same bit sequence to any number in the same group results in a number satisfying the constraint also.
For example, x = 4, y = 2. both of 01011 and 10011 have the same last 3 bits (011). Appending a 0 to each of them, resulting 010110 and 100110, both satisfy the constraint.
Here is pseudo code:
mask = (1<<(x-1)) - 1
count[0][0] = 1
for(i = 0; i < Len-1; ++i) {
for(j = 0; j < 1<<i && j < 1<<(x-1); ++j) {
if(i<x-1 || count1Bit(j*2+1)>=y)
count[i+1][(j*2+1)&mask] += count[i][j];
if(i<x-1 || count1Bit(j*2)>=y)
count[i+1][(j*2)&mask] += count[i][j];
}
}
answer = 0
for(j = 0; j < 1<<i && j < 1<<(x-1); ++j)
answer += count[Len][j];
This algorithm assumes that Len >= x. The time complexity is O(Len*2^x).
EDIT
The count1Bit(j) function counts the number of 1 in the binary representation of j.
The only input to this algorithm are Len, x, and y. It starts from an empty binary string [length 0, group 0], and iteratively tries to append 0 and 1 until length equals to Len. It also does the grouping and counting the number of binary strings satisfying the 1-bits constraint in each group. The output of this algorithm is answer, which is the number of binary strings (numbers) satisfying the constraints.
For a binary string in group [length i, group j], appending 0 to it results in a binary string in group [length i+1, group (j*2)%(2^(x-1))]; appending 1 to it results in a binary string in group [length i+1, group (j*2+1)%(2^(x-1))].
Let count[i,j] be the number of binary strings in group [length i, group j] satisfying the 1-bits constraint. If there are at least y 1 in the binary representation of j*2, then appending 0 to each of these count[i,j] binary strings yields a binary string in group [length i+1, group (j*2)%(2^(x-1))] which also satisfies the 1-bit constraint. Therefore, we can add count[i,j] into count[i+1,(j*2)%(2^(x-1))]. The case of appending 1 is similar.
The condition i<x-1 in the above algorithm is to keep the binary strings growing when length is less than x-1.
Using the example of LEN = 6, X = 3 and Y = 2...
Build an exhaustive bit pattern generator for X bits. A simple binary counter can do this. For example, if X = 3
then a counter from 0 to 7 will generate all possible bit patterns of length 3.
The patterns are:
000
001
010
011
100
101
110
111
Verify the adjacency requirement as the patterns are built. Reject any patterns that do not qualify.
Basically this boils down to rejecting any pattern containing fewer than 2 '1' bits (Y = 2). The list prunes down to:
011
101
110
111
For each member of the pruned list, add a '1' bit and retest the first X bits. Keep the new pattern if it passes the
adjacency test. Do the same with a '0' bit. For example this step proceeds as:
1011 <== Keep
1101 <== Keep
1110 <== Keep
1111 <== Keep
0011 <== Reject
0101 <== Reject
0110 <== Keep
0111 <== Keep
Which leaves:
1011
1101
1110
1111
0110
0111
Now repeat this process until the pruned set is empty or the member lengths become LEN bits long. In the end
the only patterns left are:
111011
111101
111110
111111
110110
110111
101101
101110
101111
011011
011101
011110
011111
Count them up and you are done.
Note that you only need to test the first X bits on each iteration because all the subsequent patterns were verified in prior steps.
Considering that input values are variable and wanted to see the actual output, I used recursive algorithm to determine all combinations of 0 and 1 for a given length :
private static void BinaryNumberWithOnes(int n, int dump, int ones, string s = "")
{
if (n == 0)
{
if (BinaryWithoutDumpCountContainsnumberOfOnes(s, dump,ones))
Console.WriteLine(s);
return;
}
BinaryNumberWithOnes(n - 1, dump, ones, s + "0");
BinaryNumberWithOnes(n - 1, dump, ones, s + "1");
}
and BinaryWithoutDumpCountContainsnumberOfOnes to determine if the binary number meets the criteria
private static bool BinaryWithoutDumpCountContainsnumberOfOnes(string binaryNumber, int dump, int ones)
{
int current = 0;
int count = binaryNumber.Length;
while(current +dump < count)
{
var fail = binaryNumber.Remove(current, dump).Replace("0", "").Length < ones;
if (fail)
{
return false;
}
current++;
}
return true;
}
Calling BinaryNumberWithOnes(6, 3, 2) will output all binary numbers that match
010011
011011
011111
100011
100101
100111
101011
101101
101111
110011
110101
110110
110111
111011
111101
111110
111111
Sounds like a nested for loop would do the trick. Pseudocode (not tested).
value = '0101010111110101010111' // change this line to format you would need
for (i = 0; i < (Len-x); i++) { // loop over value from left to right
kount = 0
for (j = i; j < (i+x); j++) { // count '1' bits in the next 'x' bits
kount += value[j] // add 0 or 1
if kount >= y then return success
}
}
return fail
The naive approach would be a tree-recursive algorithm.
Our recursive method would slowly build the number up, e.g. it would start at xxxxxx, return the sum of a call with 1xxxxx and 0xxxxx, which themselves will return the sum of a call with 10, 11 and 00, 01, etc. except if the x/y conditions are NOT satisfied for the string it would build by calling itself it does NOT go down that path, and if you are at a terminal condition (built a number of the correct length) you return 1. (note that since we're building the string up from left to right, you don't have to check x/y for the entire string, just also considering the newly added digit!)
By returning a sum over all calls then all of the returned 1s will pool together and be returned by the initial call, equalling the number of constructed strings.
No idea what the big O notation for time complexity is for this one, it could be as bad as O(2^n)*O(checking x/y conditions) but it will prune lots of branches off the tree in most cases.
UPDATE: One insight I had is that all branches of the recursive tree can be 'merged' if they have identical last x digits so far, because then the same checks would be applied to all digits hereafter so you may as well double them up and save a lot of work. This now requires building the tree explicitly instead of implicitly via recursive calls, and maybe some kind of hashing scheme to detect when branches have identical x endings, but for large length it would provide a huge speedup.
My approach is to start by getting the all binary numbers with the minimum number of 1's, which is easy enough, you just get every unique permutation of a binary number of length x with y 1's, and cycle each unique permutation "Len" times. By flipping the 0 bits of these seeds in every combination possible, we are guaranteed to iterate over all of the binary numbers that fit the criteria.
from itertools import permutations, cycle, combinations
def uniq(x):
d = {}
for i in x:
d[i]=1
return d.keys()
def findn( l, x, y ):
window = []
for i in xrange(y):
window.append(1)
for i in xrange(x-y):
window.append(0)
perms = uniq(permutations(window))
seeds=[]
for p in perms:
pr = cycle(p)
seeds.append([ pr.next() for i in xrange(l) ]) ###a seed is a binary number fitting the criteria with minimum 1 bits
bin_numbers=[]
for seed in seeds:
if seed in bin_numbers: continue
indexes = [ i for i, x in enumerate(seed) if x == 0] ### get indexes of 0 "bits"
exit = False
for i in xrange(len(indexes)+1):
if( exit ): break
for combo in combinations(indexes, i): ### combinatorically flipping the zero bits in the seed
new_num = seed[:]
for index in combo: new_num[index]+=1
if new_num in bin_numbers:
### if our new binary number has been seen before
### we can break out since we are doing a depth first traversal
exit=True
break
else:
bin_numbers.append(new_num)
print len(bin_numbers)
findn(6,3,2)
Growth of this approach is definitely exponential, but I thought I'd share my approach in case it helps someone else get to a lower complexity solution...
Set some condition and introduce simple help variable.
L = 6, x = 3 , y = 2 introduce d = x - y = 1
Condition: if the list of the next number hypotetical value and the previous x - 1 elements values has a number of 0-digits > d next number concrete value must be 1, otherwise add two brances with both 1 and 0 as concrete value.
Start: check(Condition) => both 0,1 due to number of total zeros in the 0-count check.
Empty => add 0 and 1
Step 1:Check(Condition)
0 (number of next value if 0 and previous x - 1 zeros > d(=1)) -> add 1 to sequence
1 -> add both 0,1 in two different branches
Step 2: check(Condition)
01 -> add 1
10 -> add 1
11 -> add 0,1 in two different branches
Step 3:
011 -> add 0,1 in two branches
101 -> add 1 (the next value if 0 and prev x-1 seq would be 010, so we prune and set only 1)
110 -> add 1
111 -> add 0,1
Step 4:
0110 -> obviously 1
0111 -> both 0,1
1011 -> both 0,1
1101 -> 1
1110 -> 1
1111 -> 0,1
Step 5:
01101 -> 1
01110 -> 1
01111 -> 0,1
10110 -> 1
10111 -> 0,1
11011 -> 0,1
11101 -> 1
11110 -> 1
11111 -> 0,1
Step 6 (Finish):
011011
011101
011110
011111
101101
101110
101111
110110
110111
111011
111101
111110
111111
Now count. I've tested for L = 6, x = 4 and y = 2 too, but consider to check the algorithm for special cases and extended cases.
Note: I'm pretty sure some algorithm with Disposition Theory bases should be a really massive improvement of my algorithm.
So in a series of Len binary digits, you are looking for a x-long segment that contains y 1's ..
See the execution: http://ideone.com/xuaWaK
Here's my Algorithm in Java:
import java.util.*;
import java.lang.*;
class Main
{
public static ArrayList<String> solve (String input, int x, int y)
{
int s = 0;
ArrayList<String> matches = new ArrayList<String>();
String segment = null;
for (int i=0; i<(input.length()-x); i++)
{
s = 0;
segment = input.substring(i,(i+x));
System.out.print(" i: "+i+" ");
for (char c : segment.toCharArray())
{
System.out.print("*");
if (c == '1')
{
s = s + 1;
}
}
if (s == y)
{
matches.add(segment);
}
System.out.println();
}
return matches;
}
public static void main (String [] args)
{
String input = "011010101001101110110110101010111011010101000110010";
int x = 6;
int y = 4;
ArrayList<String> matches = null;
matches = solve (input, x, y);
for (String match : matches)
{
System.out.println(" > "+match);
}
System.out.println(" Number of matches is " + matches.size());
}
}
The number of patterns of length X that contain at least Y 1 bits is countable. For the case x == y we know there is exactly one pattern of the 2^x possible patterns that meets the criteria. For smaller y we need to sum up the number of patterns which have excess 1 bits and the number of patterns that have exactly y bits.
choose(n, k) = n! / k! (n - k)!
numPatterns(x, y) {
total = 0
for (int j = x; j >= y; j--)
total += choose(x, j)
return total
}
For example :
X = 4, Y = 4 : 1 pattern
X = 4, Y = 3 : 1 + 4 = 5 patterns
X = 4, Y = 2 : 1 + 4 + 6 = 11 patterns
X = 4, Y = 1 : 1 + 4 + 6 + 4 = 15 patterns
X = 4, Y = 0 : 1 + 4 + 6 + 4 + 1 = 16
(all possible patterns have at least 0 1 bits)
So let M be the number of X length patterns that meet the Y criteria. Now, that X length pattern is a subset of N bits. There are (N - x + 1) "window" positions for the sub pattern, and 2^N total patterns possible. If we start with any of our M patterns, we know that appending a 1 to the right and shifting to the next window will result in one of our known M patterns. The question is, how many of the M patterns can we add a 0 to, shift right, and still have a valid pattern in M?
Since we are adding a zero, we have to be either shifting away from a zero, or we have to already be in an M where we have an excess of 1 bits. To flip that around, we can ask how many of the M patterns have exactly Y bits and start with a 1. Which is the same as "how many patterns of length X-1 have Y-1 bits", which we know how to answer:
shiftablePatternCount = M - choose(X-1, Y-1)
So starting with M possibilities, we are going to increase by shiftablePatternCount when we slide to the right. All patterns in the new window are in the set of M, with some patterns now duplicated. We are going to shift a number of times to fill up N by (N - X), each time increasing the count by shiftablePatternCount, so the full answer should be :
totalCountOfMatchingPatterns = M + (N - X)*shiftablePatternCount
edit - realized a mistake. I need to count the duplicates of the shiftable patterns that are generated. I think that's doable. (draft still)
I am not sure about my answer but here is my view.just take a look at it,
Len=4,
x=3,
y=2.
i just took out two patterns,cause pattern must contain at least y's 1.
X 1 1 X
1 X 1 X
X - represent don't care
now count for 1st expression is 2 1 1 2 =4
and for 2nd expression 1 2 1 2 =4
but 2 pattern is common between both so minus 2..so there will be total 6 pair which satisfy the condition.
I happen to be using a algoritem similar to your problem, trying to find a way to improve it, I found your question. So I will share
static int GetCount(int length, int oneBits){
int result = 0;
double count = Math.Pow(2, length);
for (int i = 1; i <= count - 1; i++)
{
string str = Convert.ToString(i, 2).PadLeft(length, '0');
if (str.ToCharArray().Count(c => c == '1') == oneBits)
{
result++;
}
}
return result;
}
not very efficent I think, but elegent solution.