How to optimize random sort algorithm? - c++

Question

Here is some random sort program I wrote in C++. It works pretty fine for 10 elements or so. But for 15 elements it works so slow I can't even wait enough to get the result. Is there some way to optimize random sort algorithm?
Here's my code:
// randomsort.h
#ifndef RANDOMSORT_H
#define RANDOMSORT_H
#include <stdlib.h>
#include <time.h>
class RandomSort
{
private:
template <class T>
static bool isOrdered(T*, int);
public:
template <class T>
static int sort(T*, int);
};
template <class T>
bool RandomSort::isOrdered(T* arr, int size)
{
for(int i = 1; i < size; i++)
{
if(arr[i-1] > arr[i])
{
return false;
}
}
return true;
}
template <class T>
int RandomSort::sort(T* arr, int size)
{
int stepAmount = 0;
srand(time(NULL));
while(!isOrdered(arr, size))
{
int i = rand() % size;
int j = rand() % size;
std::swap(arr[i], arr[j]);
stepAmount++;
}
return stepAmount;
}
#endif // RANDOMSORT_H
And main.cpp file
// main.cpp
#include <iostream>
#include "randomsort.h"
int main()
{
int size;
std::cout << "Enter amount of elements to sort: ";
std::cin >> size;
std::cout << std::endl;
int arr[size];
for(int i = 0; i < size; i++)
{
arr[i] = (rand() % (size * 10));
}
std::cout << "Input array: " << std::endl;
for(int i = 0; i < size; i++)
std::cout << arr[i] << ' ';
std::cout << std::endl << std::endl;
int stepAmount = RandomSort::sort(arr, size);
std::cout << "Output array: " << std::endl;
for(int i = 0; i < size; i++)
std::cout << arr[i] << ' ';
std::cout << std::endl << std::endl;
std::cout << "Number of steps: " << stepAmount;
return 0;
}
Any suggestions?

Answer 1

Your code is completely random. So it can swap when it should not. An easy fix would be to swap only if you need it.
int i = rand() % size;
int j = rand() % size;
// to know which should be first
if (i > j)
std::swap(i, j);
if (arr[i] > arr[j])
std::swap(arr[i], arr[j]);
Your array probably will not be sorted immediately, so you could also test if it is sorted only every five steps (for example) instead of every step.
But i think the most important is, you should not expect good performances from such an algorithm.