//#define _GLIBCXX_DEBUG
//#pragma GCC optimize("Ofast")
//#pragma GCC optimize("unroll-loops")
//#pragma GCC target("sse,sse2,sse3,ssse3,sse4,popcnt,abm,mmx,avx,tune=native")
#include <bits/stdc++.h>
using namespace std;
//#include <ext/pb_ds/assoc_container.hpp>
//
//using namespace __gnu_pbds;
//
//template<typename T>
//using ordered_set = tree<T, null_type, less < T>, rb_tree_tag, tree_order_statistics_node_update>;
//template<typename T>
//using normal_queue = priority_queue<T, vector<T>, greater<>>;
mt19937 rnd(chrono::steady_clock::now().time_since_epoch().count());
#define trace(x) cout << #x << ": " << (x) << endl;
#define all(x) begin(x), end(x)
#define rall(x) rbegin(x), rend(x)
#define uniq(x) x.resize(unique(all(x)) - begin(x))
#define sz(s) ((int) size(s))
#define pii pair<int, int>
#define mp(x, y) make_pair(x, y)
#define int128 __int128
#define pb push_back
#define popb pop_back
#define eb emplace_back
#define fi first
#define se second
#define itn int
typedef long long ll;
typedef pair<ll, ll> pll;
typedef long double ld;
typedef double db;
typedef unsigned int uint;
template<typename T>
bool ckmn(T &x, T y) {
if (x > y) {
x = y;
return true;
}
return false;
}
template<typename T>
bool ckmx(T &x, T y) {
if (x < y) {
x = y;
return true;
}
return false;
}
int bit(int x, int b) {
return (x >> b) & 1;
}
int rand(int l, int r) { return (int) ((ll) rnd() % (r - l + 1)) + l; }
const ll infL = 3e18;
const int infI = 1000000000 + 7;
const int infM = 0x3f3f3f3f; //a little bigger than 1e9
const ll infML = 0x3f3f3f3f3f3f3f3fLL; //4.5e18
const int N = 100002;
const int mod = 998244353;
const ld eps = 1e-9;
struct cht_min {
struct line {
ll k, b;
int id;
double f;
line() = default;
line(ll a, ll B, int gg) {
k = a, b = B, id = gg;
}
double intersect(line now) const {
return (ld) (b - now.b) / (now.k - k);
}
ll get(ll x) const {
return k * x + b;
}
};
vector<line> d;
int idx = 0;
void add(line now) {
while (d.size() > 1) {
if (d.back().k == now.k) {
if (d.back().b < now.b) {
return;
}
}
if (now.intersect(d.back()) > d.back().f) break;
d.pop_back();
}
if (!d.empty())
now.f = now.intersect(d.back());
else
now.f = -infL;
d.push_back(now);
}
pair<ll, int> get(ll x) {
if (d.empty()) return {infL, -1};
idx = min(idx, (int) d.size() - 1);
while (idx < (int) d.size() - 1 && d[idx].get(x) > d[idx + 1].get(x))
++idx;
return {d[idx].get(x), d[idx].id};
}
};
int pref[N];
const int K = 201;
ll dp[2][N];
int p[K][N];
int main() {
ios::sync_with_stdio(false);
cin.tie(nullptr);
int n, k;
cin >> n >> k;
for (int i = 1; i <= n; ++i) {
cin >> pref[i];
pref[i] += pref[i - 1];
}
//we need to minimize the squares of sums in each group
for (int i = 1; i <= n; ++i) {
dp[0][i] = pref[i] * 1LL * pref[i];
}
memset(p, -1, sizeof(p));
for (int cut = 1; cut <= k; ++cut) {
cht_min t;
int nw = cut & 1;
int pr = nw ^ 1;
fill(dp[nw], dp[nw] + n + 1, infL);
for (int i = 1; i <= cut; ++i) {
t.add({-2 * pref[i], pref[i] * 1LL * pref[i] + dp[pr][i], i});
}
for (int i = cut + 1; i <= n; ++i) {
auto[ff, ss] = t.get(pref[i]);
if (ff < infL) {
ff += pref[i] * 1LL * pref[i];
if (ckmn(dp[nw][i], ff)) {
p[cut][i] = ss;
}
t.add({-2 * pref[i], pref[i] * 1LL * pref[i] + dp[pr][i], i});
}
}
}
vector<int> ans;
for (int i = n, cut = k; i != -1;) {
ans.push_back(i);
i = p[cut--][i];
}
reverse(all(ans));
ans.pop_back();
ll val = pref[n] * 1LL * pref[n] - dp[k & 1][n];
cout << val / 2 << '\n';
for (int x: ans) cout << x << " ";
return 0;
}
