While doing some backtracking exercises, got stuck here:
Write a recursive algorithm that generates all partitions of a given n numbers. Of the partitions that differ only in the order of the members, we need to list only one, the last from a lexicographic point of view. Write the solutions lexicographically in ascending order. 1 <= n <= 100
n = 5 would be:
1 1 1 1 1
2 1 1 1
2 2 1
3 1 1
3 2
4 1
5
I searched over the web for solutions and found this, but it's not quite right and I don't know how to perfect it. First of all it's not in lexicographical order, and doesn't include the actual number.
So for 5 it outputs:
1 1 1 1 1
1 1 1 2
1 1 3
1 2 2
1 4
2 3
Here is my code, where I tried to correct it, it's better, but still not exactly the way the example is..
#include <iostream>
#include <vector>
using namespace std;
void print(const vector<vector<int>>& output)
{
for(int i = 0; i < output.size(); i){
for(int j = output[i].size()-1; j >= 0; --j){
cout << output[i][j] << " ";
}
cout << endl;
}
}
void iteration(int n, int current_sum, int start, vector<vector<int>>& output, vector<int>& result)
{
if (n == current_sum) {
output.push_back(result);
}
for (int i = start; i < n; i ) {
int temp = current_sum i;
if (temp <= n) {
result.push_back(i);
iteration(n, temp, i, output, result);
result.pop_back();
}
else {
return ;
}
}
}
void decompose(int n)
{
vector<vector<int>> output;
vector<int> result;
iteration(n, 0, 1, output, result);
print(output);
cout << n << endl;
return ;
}
int main()
{
int n = 5;
decompose(n);
return 0;
}
So, now the output is:
1 1 1 1 1
2 1 1 1
3 1 1
2 2 1
4 1
3 2
5
So, the 2 2 1 and 3 2 is in the wrong place.. And the bigger the "n" number is, the bigger the mess..
Can someone help?
CodePudding user response:
If you just want the answer ... here is some code that I translated from Python to C# a few years ago and just translated to C for this answer.
#include <iostream>
#include <vector>
std::vector<int> join(int val, const std::vector<int>& vec) {
std::vector<int> joined;
joined.reserve(vec.size() 1);
joined.push_back(val);
std::copy(vec.begin(), vec.end(), std::back_inserter(joined));
return joined;
}
std::vector<int> tail(const std::vector<int>& v) {
std::vector<int> t(v.size() - 1);
std::copy(std::next(v.begin()), v.end(), t.begin());
return t;
}
std::vector<std::vector<int>> partitions(int n) {
if (n == 0) {
return { {} };
}
auto partitions_n_minus_1 = partitions(n - 1);
std::vector<std::vector<int>> partitions_n;
for (const auto& p: partitions_n_minus_1) {
partitions_n.push_back( join({ 1 }, p) );
if (!p.empty() && (p.size() < 2 || p[1] > p[0])) {
partitions_n.push_back(join({ p[0] 1 }, tail(p)));
}
}
return partitions_n;
}
int main()
{
auto partions = partitions( 6 );
for (const auto& p : partions) {
for (int i : p) {
std::cout << i << " ";
}
std::cout << "\n";
}
return 0;
}
The Python code this comes from is here, Generator for integer partitions (Python recipe), with some explanation by the author in comments.
In my C# code I was able to retain the generator/yield idiom used in the Python version but I'm afraid I currently do not know how to use C coroutines yet so I've moved it to a standard function implementation which will be less efficient as it requires a lot of copying.
CodePudding user response:
When you get more practice programming, you'll learn how to keep things simple:
#include <iostream>
#include <vector>
void decompose(int n, std::vector<int> prefix = {})
{
if (n == 0) {
for (int a : prefix) { std::cout << a << ' '; }
std::cout << std::endl;
}
else {
int max = prefix.size() ? std::min(prefix.back(), n) : n;
prefix.push_back(1);
for (int i = 1; i <= max; i ) {
prefix.back() = i;
decompose(n - i, prefix);
}
}
}
int main()
{
decompose(5);
}