Print 2-D Array in clockwise expanding spiral from center - c++

I have an guaranteed to be a perfect square matrix. I want to start at the center of the matrix in this case it would be matrix[2][2], I know how to figure the center (int)(dimensions / 2). I need to output the contents of the array in this following outward spiral pattern. Of course the algorithm should work with any perfect square matrix. I wasn't sure if this algorithm already existed and I didn't want to re-invent the wheel.
int dimensions / 2;
21 22 23 24 25
20 7 8 9 10
19 6 1 2 11
18 5 4 3 12
17 16 15 14 13
The output for this example should be
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

Let's identify the patterns first ..
Even-Size Square Matrix, Example: 4x4
07 > 08 > 09 > 10
^ v
06 (01)> 02 11
^ v v
05 < 04 < 03 12
v
[16]< 15 < 14 < 13
Starting Point: [2, 2] ~ [SIZE/2, SIZE/2]
Ending Point: [4, 1] ~ [SIZE, 1]
Chains: Count(K-chain)=2 for K = 1..(SIZE-2)
+ 3 for Count = SIZE-1
Odd-Size Square Matrix, Example: 5x5
21 > 22 > 23 > 24 >[25]
^
20 07 > 08 > 09 > 10
^ ^ v
19 06 (01)> 02 11
^ ^ v v
18 05 < 04 < 03 12
^ v
17 < 16 < 15 < 14 < 13
Starting Point: [2, 2] ~ [⌊SIZE/2⌋, ⌊SIZE/2⌋]
Ending Point: [1, 5] ~ [1, SIZE]
Chains: Count(K-chain)=2 for K = 1..(SIZE-2)
+ 3 for Count = SIZE-1
Live Code
void print_spiral (int ** matrix, int size)
{
int x = 0; // current position; x
int y = 0; // current position; y
int d = 0; // current direction; 0=RIGHT, 1=DOWN, 2=LEFT, 3=UP
int c = 0; // counter
int s = 1; // chain size
// starting point
x = ((int)floor(size/2.0))-1;
y = ((int)floor(size/2.0))-1;
for (int k=1; k<=(size-1); k++)
{
for (int j=0; j<(k<(size-1)?2:3); j++)
{
for (int i=0; i<s; i++)
{
std::cout << matrix[x][y] << " ";
c++;
switch (d)
{
case 0: y = y + 1; break;
case 1: x = x + 1; break;
case 2: y = y - 1; break;
case 3: x = x - 1; break;
}
}
d = (d+1)%4;
}
s = s + 1;
}
}

As this smells like homework then no code or direct answer instead just few hints:
You can look at this as an turtle problem:
Let m be movement by 1 cell and r be rotation by 90 degrees in your spiral direction (CW or CCW). then the spiral can be encoded as series of turtle commands forming this pattern (from the start point):
m,r,m,r,
m,m,r,m,m,r,
m,m,m,r,m,m,m,r,
m,m,m,m,r,m,m,m,m,r,
m,m,m,m,m,r
As you can see you start with 1x move then rotate after two repetition you switch to 2x move, after 2 moves switch to 3x move,... and so on. This can be done with just few for loops (or just by one with some proper iterations and stopping when matrix number of cells is hit ... or the endpoint corner is hit)
You need to handle even/odd matrix sizes
for odd matrix sizes is the middle point easy. For even sizes it is a bit more complicated. if you use CW rotation then use the rounded down division result of halving the size and start with moving to the right. (if you need different spiral then you need to add +1 to x and/or y and change starting direction) so the spiral will stay centered.
So If you got even sized matrix then the last movement is twice if you got odd size then the last movement is there just once (like in this example)
rotation
Have direction stored as 2D vector. for example d=(+1,0) means right. to rotate 2D vector you just swap coordinates and negate one axis (which one means the CW/CCW). For example (x,y) -> (y,-x)
movement
Store current position as 2D vector too. The movement is just adding current direction vector to it.
Have fun with solving this ...

int radius = 0;
int i = centerX;
int j = centerY;
Debug.Log("i=" + i + "; j=" + j);
++i;
radius += 2;
while ((i < dimm) && (i >= 0))
{
for (int c = j; j < c + radius; j++)
Debug.Log("i=" + i + "; j=" + j);
--j;
--i;
for (int c = i; i > c - radius + 1; i--)
Debug.Log("i=" + i + "; j=" + j);
if (i < 0)
break;
else
Debug.Log("i=" + i + "; j=" + j);
--j;
for (int c = j; j > c - radius; j--)
Debug.Log("i=" + i + "; j=" + j);
++i;
++j;
for (int c = i; i < c + radius; i++)
Debug.Log("i=" + i + "; j=" + j);
radius += 2;
}
This Code will output counterclockwise squared matrix (dimm X dimm) indexes from custom center(CenterX, CenterY); and end up, if it came out of matrix size.

bool IsIterationComplete(int iteration, int nCenter, std::vector<std::vector<bool>>& vVisited)
{
int nHigh = nCenter+iteration;
int nLow = nCenter-iteration;
//cout<<endl<<"High "<<nHigh<<"Low "<<nLow<<endl;
for(int i=nLow;i<=nHigh;i++)
{
if(!vVisited[nHigh][i] || !vVisited[nLow][i])
return false;
}
for(int i=nLow;i<=nHigh;i++)
{
if(!vVisited[i][nHigh] || !vVisited[i][nLow])
return false;
}
return true;
}
void PrintSpiral(std::vector<std::vector<int>>& vMat,std::vector<std::vector<bool>>& vVisited, int row, int col, int nCenter, int iteration)
{
cout<<vMat[row][col];
vVisited[row][col]=true;
if(row==0 && col==0)
return;
if(IsIterationComplete(iteration,nCenter,vVisited))
iteration++;
//cout<<endl<<"row "<<row<<" column "<<col<<"Iteration "<<iteration<<endl;
//left
if(col-1>=0 && !vVisited[row][col-1] && col-1>=nCenter-iteration)
{
cout<<"Left "<<endl;
PrintSpiral(vMat,vVisited,row,col-1,nCenter,iteration);
}
//Down
if((row+1)<vMat.size() && !vVisited[row+1][col] && row+1<=nCenter+iteration)
{
cout<<"Down "<<endl;
PrintSpiral(vMat,vVisited,row+1,col,nCenter,iteration);
}
//right
if((col+1)<vMat.size() && !vVisited[row][col+1] && col+1<=nCenter+iteration)
{
cout<<"Right "<<endl;
PrintSpiral(vMat,vVisited,row,col+1,nCenter,iteration);
}
//up
if(row-1>=0 && !vVisited[row-1][col] && row-1>=nCenter-iteration)
{
cout<<"Up "<<endl;
PrintSpiral(vMat,vVisited,row-1,col,nCenter,iteration);
}
}
int main (int argc, const char * argv[])
{
int nCount=1;
std::vector<std::vector<int>> vMat;
std::vector<std::vector<bool>> vVisited;
for(int i=0; i<7; i++)
{
std::vector<int> row;
std::vector<bool> visitedRow;
for(int j=0; j<7; j++)
{
row.push_back(nCount);
cout<<nCount<<'\t';
nCount++;
visitedRow.push_back(false);
}
cout<<'\n';
vMat.push_back(row);
vVisited.push_back(visitedRow);
}
cout<<'\n';
PrintSpiral(vMat,vVisited,vMat.size()/2,vMat.size()/2,vMat.size()/2,0);
return 0;
}

Related

Coursera DSA Algorithmic toolbox week 4 2nd question- Partitioning Souvenirs

Problem Statement-
You and two of your friends have just returned back home after visiting various countries. Now you would
like to evenly split all the souvenirs that all three of you bought.
Problem Description
Input Format- The first line contains an integer ... The second line contains integers v1, v2, . . . ,vn separated by spaces.
Constraints- 1 . .. . 20, 1 . .... . 30 for all ...
Output Format- Output 1, if it possible to partition 𝑣1, 𝑣2, . . . , 𝑣𝑛 into three subsets with equal sums, and
0 otherwise.
What's Wrong with this solution? I am getting wrong answer when I submit(#12/75) I am solving it using Knapsack without repetition taking SUm/3 as my Weight. I back track my solution to replace them with 0. Test cases run correctly on my PC.
Although I did it using OR logic, taking a boolean array but IDK whats wrong with this one??
Example- 11
17 59 34 57 17 23 67 1 18 2 59
(67 34 17)are replaced with 0s. So that they dont interfare in the next sum of elements (18 1 23 17 59). Coz both of them equal to 118(sum/3) Print 1.
#include <iostream>
#include <vector>
using namespace std;
int partition3(vector<int> &w, int W)
{
int N = w.size();
//using DP to find out the sum/3 that acts as the Weight for a Knapsack problem
vector<vector<int>> arr(N + 1, vector<int>(W + 1));
for (int k = 0; k <= 1; k++)
{
//This loop runs twice coz if 2x I get sum/3 then that ensures that left elements will make up sum/3 too
for (int i = 0; i < N + 1; i++)
{
for (int j = 0; j < W + 1; j++)
{
if (i == 0 || j == 0)
arr[i][j] = 0;
else if (w[i - 1] <= j)
{
arr[i][j] = ((arr[i - 1][j] > (arr[i - 1][j - w[i - 1]] + w[i - 1])) ? arr[i - 1][j] : (arr[i - 1][j - w[i - 1]] + w[i - 1]));
}
else
{
arr[i][j] = arr[i - 1][j];
}
}
}
if (arr[N][W] != W)
return 0;
else
{
//backtrack the elements that make the sum/3 and = them to 0 so that they don't contribute to the next //elements that make sum/3
int res = arr[N][W];
int wt = W;
for (int i = N; i > 0 && res > 0; i--)
{
if (res == arr[i - 1][wt])
continue;
else
{
std::cout << w[i - 1] << " ";
res = res - w[i - 1];
wt = wt - w[i - 1];
w[i - 1] = 0;
}
}
}
}
if (arr[N][W] == W)
{
return 1;
}
else
{
return 0;
}
}
int main()
{
int n;
std::cin >> n;
vector<int> A(n);
int sum = 0;
for (size_t i = 0; i < A.size(); ++i)
{
int k;
std::cin >> k;
A[i] = k;
sum += k;
}
if (sum % 3 == 0)
std::cout << partition3(A, sum / 3) << '\n';
else
std::cout << 0;
}
Sum/3 can be achieved by multiple ways!!!! So backtracking might remove a subset that has an element that should have been a part of some other subset
8 1 6 is 15 as well as 8 2 5 makes 15 so better is u check this
if(partition3(A, sum / 3) == sum / 3 && partition3(A, 2 * (sum / 3)) == 2 * sum / 3 && sum == partition3(A, sum))

Every time I change the constant, the result is different

I want to code the 01 Knapsack problem with c++. Each time I change the value of a constant MAX to a different value, such as 100,90, or 80, the result is different. What's the cause? I'm using Visual Studio 2019.
The input is a text file, for example.
20
40 35 18 4 10 2 70 20 39 37 7 5 10 8 15 21 50 40 10 30
100 50 45 20 10 5 31 10 20 19 4 3 6 8 12 7 10 2 5 5
137
The first row is the number of objects, the second row is the price, the third row is the weight, and the last row is the size of the knapsack.
#include <stdio.h>
#include <time.h>
#define MAX 100
#define CLOCKS_PER_MS CLOCKS_PER_SEC/1000
int X_d[MAX] = { 0 }; //solution vector
int max(int a, int b) {
if (a >= b)
return a;
else
return b;
}
void dynamic(int n, int M, int p[], int w[]) {
int result;
int i, y, k;
int P[MAX][MAX] = { 0 };
clock_t start, finish;
start = clock();
for (i = 0; i <= n; i++) {
for (y = 0; y <= M; y++) {
if (i == 0 || y == 0)
P[i][y] = 0;
else if (w[i - 1] > y)
P[i][y] = P[i - 1][y];
else
P[i][y] = max(P[i - 1][y], p[i - 1] + P[i - 1][y - w[i - 1]]);
}
}
finish = clock();
result = P[n][M];
y = M;
for (i = n; i > 0 && result > 0; i--) {
if (result == P[i - 1][y]) {
continue;
}
else {
X_d[i - 1] = 1;
result = result - p[i - 1];
y = y - w[i - 1];
}
}
printf("\n(1) Dynamic Programming");
printf("\nThe maximum profit is $%d", P[n][M]);
printf("\nThe solution vetor X = ( ");
for (k = 0; k < n; k++)
printf("%d ", X_d[k]);
printf(")\n");
printf("The execution time is %f milliseconds.\n", (float)(finish - start) / CLOCKS_PER_MS);
}
int main() {
int i, j;
int num, M;
int p[MAX] = { 0 }, w[MAX] = { 0 };
FILE* fp = NULL;
fopen_s(&fp, "p2data6.txt", "r");
fscanf_s(fp, "%d", &num);
for (i = 0; i < num; i++)
fscanf_s(fp, "%d", &p[i]);
for (i = 0; i < num; i++)
fscanf_s(fp, "%d", &w[i]);
fscanf_s(fp, "%d", &M);
printf("n = %d\n", num);
printf("pi = ");
for (i = 0; i < num; i++)
printf("%3d ", p[i]);
printf("\nwi = ");
for (i = 0; i < num; i++)
printf("%3d ", w[i]);
printf("\npi/wi = ");
for (i = 0; i < num; i++)
printf("%f ", (double)p[i] / w[i]);
printf("\nM = %d\n", M);
dynamic(num, M, p, w);
return 0;
}
Geeks for Geeks has a very good explanation of this problem here:
To quote the Geeks for Geeks article by Sam007:
A simple solution is to consider all subsets of items and calculate the total weight and value of all subsets. Consider the only subsets whose total weight is smaller than W. From all such subsets, pick the maximum value subset.
1) Optimal Substructure:
To consider all subsets of items, there can be two cases for every item: (1) the item is included in the optimal subset, (2) not included in the optimal set.
Therefore, the maximum value that can be obtained from n items is max of following two values.
1) Maximum value obtained by n-1 items and W weight (excluding nth item).
2) Value of nth item plus maximum value obtained by n-1 items and W minus weight of the nth item (including nth item).
If weight of nth item is greater than W, then the nth item cannot be included and case 1> is the only possibility.
2) Overlapping Subproblems
Following is recursive implementation that simply follows the recursive structure mentioned above.

Fill Matrix in a special order C++

I want to fill a 8 x 8 matrix with values in a special order (see example below), but I don´t know how to do that. Each numer stands for the ordering number: For example: #3 in the matrix is the third value of a e.g. a measurment I want to add.
The Order should be:
1 2 5 6 17 18 21 22
3 4 7 8 19 20 23 24
9 10 13 14 25 26 29 30
11 12 15 16 27 28 31 32
33 34 37 38 49 50 53 54
35 36 39 40 51 52 55 56
41 42 45 46 57 58 61 62
43 44 47 48 59 60 63 64
Does anybody knows an algorithmus to do that?
I have tried this, but that´s not a good way to to it, and it´s not working for the whole matrix
int b= 0, ii = 0, a = 0, iii = 0
i are different measurement values
and now a for loop
if (ii == 1)
{
b++;
}
if (ii == 2)
{
a++, b--;
}
if (ii == 3)
{
b ++;
}
tempMatrix[a][b] = i;
cout << "TempMatrix " << tempMatrix[a][b] << " a " << a << " b " << b << endl;
if (ii == 3)
{
ii = -1;
a --;
b ++;
}
if (iii == 7)
{
a = a + 2;
b = 0;
iii = -1;
}
Use recursion:
#include <iostream>
using namespace std;
void f(int a[8][8], int current, int x, int y, int size) {
if (size == 1) {
a[x][y] = current;
return;
} else {
size /= 2;
int add_for_each_square = size * size;
f(a, current, x, y, size);
f(a, current + add_for_each_square, x, y + size, size);
f(a, current + 2 * add_for_each_square, x + size, y, size);
f(a, current + 3 * add_for_each_square, x + size, y + size, size);
}
}
int main() {
int a[8][8];
f(a, 1, 0, 0, 8);
for (int i = 0; i < 8; ++i) {
for (int j = 0; j < 8; ++j) {
cout << a[i][j] << " ";
}
cout << endl;
}
}
If the matrix will always be a fixed size, then you can generate two lookup tables for row and column indexes into the matrix. Then, just pass your index through these tables to get the desired positions in the matrix.
const auto MATRIX_SIZE = 8;
const std::array<int, MATRIX_SIZE*MATRIX_SIZE> row_lookup = {{...}}; //put pre-computed values here.
const std::array<int, MATRIX_SIZE*MATRIX_SIZE> col_lookup = {{...}};
for(size_t i = 0; i < MATRIX_SIZE * MATRIX_SIZE; i++)
{
auto val = get_coefficient(i);
auto row = row_lookup[i];
auto col = col_lookup[i];
mat[col][row] = val;
}

generate a 2d array of integers from given sums of its rows and columns

I want to generate an array of integers where the total sum of each row and column in the array is known , for example if I create a 4 by 4 array in c++ and then populate it pseudo randomly with numbers between 1 and 100:
int array[4][4] = {} ;
for(int x = 0 ; x<4 ; x++){
for(int y = 0 ; y<4 ; y++){
array[x][y] = rand() % 100 + 1 ;
}
}
the array would be :
8, 50, 74, 59
31, 73, 45, 79
24, 10, 41, 66
93, 43, 88, 4
then if I sum each row and each column by :
int rowSum[4] = {} ;
int columnSum[4] = {} ;
for(int x = 0 ; x < 4; x++){
for(int y = 0 ; y < 4; y++){
rowSum[x] += array[x][y] ;
columnSum[y] += array[x][y] ;
}
}
the rowSum would be {191,228,141,228} and the columnSum = {156,176,248,208}
what I'm trying to do at this point is to generate any random 4x4 1~100 array that will satisfy rowSum and columnSum I understand there is thousands of different arrays that will sum up to the same row and column sum ,and I've been trying to write the part of the code that will generate it , I would really appreciate it if anyone can give me a clue .
It is very easy to find some solution.
Start with generating row that sum to given values. It could be as simple as making all values in each row approximately equal to rowSum[i]/n, give or take one. Of course sums of columns will not match at this point.
Now fix the columns from the leftmost to the rightmost. To fix i th column, distribute the difference between the desired sum and the actual sum equally between column entries, and then fix each row by distributing the added value equally between items i+1...n of the row.
It is easier done than said:
void reconstruct (int array[4][4], int rows[4], int cols[4])
{
// build an array with each row adding up to the correct row sum
for (int x = 0; x < 4; x++){
int s = rows[x];
for(int y = 0; y < 4 ; y++){
array[x][y] = s / (4 - y);
s -= array[x][y];
}
}
// adjust columns
for(int y = 0; y < 4 ; y++){
// calculate the adjustment
int s = 0;
for (int x = 0; x < 4; x++){
s += array[x][y];
}
int diff = s - cols[y];
// adjust the column by diff
for (int x = 0; x < 4; x++){
int k = diff / (4 - x);
array[x][y] -= k;
diff -= k;
// adjust the row by k
for (int yy = y + 1; yy < 4; ++yy)
{
int corr = k / (4 - yy);
array[x][yy] += corr;
k -= corr;
}
}
}
}
This array won't be random of course. One can randomise it by selecting x1, x2, y1, y2 and d at random and executing:
array[x1][y1] += d
array[x1][y2] -= d
array[x2][y1] -= d
array[x2][y2] += d
taking care that the resulting values won't spill out of the desired range.
Here's the quick and dirty brute force search mentioned in comments. It ought to give you a starting point. This is C, not C++.
You never said it, but I'm assuming you want the matrix elements to be non-negative. Consequently, this searches the space where each element a[i][j] can have any value in [0..min(rowsum[i], colsum[j])] with the search cut off when assigning the next array element value would admit no possible future solution.
#include <stdio.h>
int a[4][4] = {
{-1, -1, -1, -1},
{-1, -1, -1, -1},
{-1, -1, -1, -1},
{-1, -1, -1, -1}};
int rs[] = {191, 228, 141, 228};
int cs[] = {156, 176, 248, 208};
long long n_solutions = 0;
void research(int i, int j, int ii, int jj, int val);
void print_a(void);
void search(int i, int j) {
if (j < 3) {
if (i < 3) {
int m = rs[i] < cs[j] ? rs[i] : cs[j];
for (int val = 0; val <= m; ++val) research(i, j, i, j + 1, val);
} else {
if (rs[3] >= cs[j]) research(i, j, i, j + 1, cs[j]);
}
} else {
if (i < 3) {
if (cs[j] >= rs[i]) research(i, 3, i + 1, 0, rs[i]);
} else {
if (rs[3] == cs[3]) {
a[3][3] = rs[i];
if (++n_solutions % 100000000 == 0) {
printf("\n%lld\n", n_solutions);
print_a();
}
a[3][3] = -1;
}
}
}
}
void research(int i, int j, int ii, int jj, int val) {
a[i][j] = val; rs[i] -= val; cs[j] -= val;
search(ii, jj);
rs[i] += val; cs[j] += val; a[i][j] = -1;
}
void print_a(void) {
for (int i = 0; i < 4; ++i) {
for (int j = 0; j < 4; ++j)
printf("%4d", a[i][j]);
printf("\n");
}
}
int main(void) {
search(0, 0);
printf("Total solutions: %lld\n", n_solutions);
return 0;
}
For example, if you replace the simple for loop with this, you won't get so many zeros in the upper left hand corner:
int b = m / 2; // m/2 can be replaced with any int in [0..m], e.g. a random value.
research(i, j, i, j + 1, b);
for (int d = 1; b + d <= m || b - d >= 0; ++d) {
if (b + d <= m) research(i, j, i, j + 1, b + d);
if (b - d >= 0) research(i, j, i, j + 1, b - d);
}
Here's the 2-billionth solution:
78 56 28 29
39 20 84 85
28 34 61 18
11 66 75 76
The problem becomes interesting if we place condition that the matrix elements must be non-negative integers. Here's an O(mn) JAVA solution based on greedy algorithm.
int m=rowSum.length;
int n=colSum.length;
int mat[][] = new int[m][n];
for(int i=0;i<m;i++){
for(int j=0;j<n;j++){
int tmp=Math.min(rowSum[i],colSum[j]);
mat[i][j]=tmp;
rowSum[i]-=tmp;
colSum[j]-=tmp;
}
}
return mat;

Debugging a histogram in C++

I want to build a histogram with ell intervals. The size of each interval is computed as k = ceil(m/ell) where m is the maximum number in the data set. That is, the interval i should be [(i - 1) * k, i * k). If a data set is given by the numbers 16 33 55 57 8 47 1 21 14 73 6 59 29 57 20 95 77 5 62 48 and the number of intervals is ell = 10, the histogram has to be textually represented as
0: 4
10: 2
20: 3
30: 1
40: 2
50: 4
60: 1
70: 2
80: 0
90: 1
I need to write a program, that reads in the following values (from the standard input cin) :
the number ell of intervals
the size n of the data set
n non-negative integers
Here is my code so far:
#include<iostream>
#include<math.h>
#include<algorithm>
using namespace std;
int main() {
double ell;
int n; // size of the data set
double m = 0;
int * a;
int * x;
int * y;
cin >> ell;
cin >> n;
a = new int[n];
// finding max element in array a
for (int i = 0; i < n; i++) {
cin >> a[i];
if (a[i] > m) {
m = a[i];
}
}
int k = ceil(m/ell);
x = new int[(int)(ell)];
y = new int[(int)(ell)];
for (int i = 1; i <= ell; i++) {
x[i] = (i - 1) * k;
for (int j = 0; j < n; j++) {
if (a[j] >= (i - 1) * k && a[j] < i * k && x[i] != (ell - 1) * k) {
y[i] += 1;
} else if (a[j] >= (ell - 1) * k && x[i] == (ell - 1) * k) {
y[i] += 1;
} else {
y[i] += 0;
}
}
cout << x[i] << ": " << y[i] << endl;
}
delete [] a;
delete [] x;
delete [] y;
return 0;
}
If I input this:
10 20 16 33 55 57 8 47 1 21 14 73 6 59 29 57 20 95 77 5 62 48
I get
0: 4
10: 2
20: 3
30: 1
40: 2
50: 4
60: 1
70: 2
80: 0
90: 1
But if I input the same many times, I will get a weird output like
0: 4
10: 268501820
20: 1073741827
30: 112
40: 2
50: 4
60: 1
70: 2
80: 0
90: 6
Why is this happening and how can I fix my problem?
x = new int[(int)(ell)];
y = new int[(int)(ell)];
Creates objects on the heap, but unless the objects (i.e. int) have a constructor, then the memory is left uninitialized.
The array x is correctly initialized, but y is left (expected to be 0).
So the fix is ...
x = new int[(int)(ell)];
y = new int[(int)(ell)];
for( int i = 0; i < ell ; i++ ){
x[i] = 0; // tidier to do it obviously.
y[i] = 0; // essential to get y[i] += 1; to be meaningful.
}
You never initialise your memory. You create dynamic arrays with both x and y but you are not guaranteed that the memory is zeroed. You get lucky for the most part that but when there are values left in the memory your program uses you get the outliers that you see. Put something like this in and it should fix your issues.
for(int m = 0;m < (int)ell; m++) {
x[m] = 0;
y[m] = 0;
}