[题解]HDU5634 Rikka with Phi
#1.0 题意简述
维护数列,三种操作:
- 给定 \(l,r\),将 \(a_i(i\in[l,r])\) 变为 \(\varphi(a_i)\);
- 给定 \(l,r,x\),将 \(a_i(i\in[l,r])\) 变为 \(x\);
- 询问区间和;
#2.0 大体思路
发现所有的数都有归于统一(五桶!)的趋势,一个数在进行不超过 \(\log n\) 次操作后会变为 \(1\),但是区间取 \(\varphi\) 的操作显然不能直接区间维护,于是只能暴力修改,但是,注意到如果整个区间的数都相同,那么可以直接区间取 \(\varphi\),于是考虑维护一个区间是否相同的标记,如果相同,那么就直接区间取 \(\varphi\) 就可以了,如果没有区间覆盖的操作,那么时间复杂度最差为 \(O(n\log^2n)\),即数各不相同,每次单点取 \(\varphi\),直到全部变为 \(1\)。区间覆盖的存在不会影响整体的时间复杂度。
注意区间相同标记是自底向上维护和自顶向下维护并存。
#3.0 Code
const int N = 700010;
const int M = 10000010;
const int INF = 0x3fffffff;
struct Node {
int ls, rs; ll sum, tag;
inline Node() {ls = rs = sum = tag = 0;}
inline void del() {ls = rs = sum = tag = 0;}
} p[N];
int T, n, m, cnt, phi[M], prm[N];
int rt, pcnt, nprm[M]; ll a[N];
void Get_Phi(int x) {
phi[1] = 1;
for (int i = 2; i <= x; ++ i) {
if (!nprm[i]) prm[++ pcnt] = i, phi[i] = i - 1;
for (int j = 1; j <= pcnt; ++ j) {
if (prm[j] * i > x) {break;} nprm[prm[j] * i] = true;
if (i % prm[j]) phi[prm[j] * i] = phi[i] * (prm[j] - 1);
else {phi[prm[j] * i] = phi[i] * prm[j]; break;}
}
}
}
inline void clear() {
for (int i = 1; i <= cnt; ++ i) p[i].del();
cnt = 0;
}
inline void pushup(int k) {
int ls = p[k].ls, rs = p[k].rs;
p[k].sum = p[ls].sum + p[rs].sum;
if (p[ls].tag && p[ls].tag == p[rs].tag)
p[k].tag = p[ls].tag;
else p[k].tag = 0;
}
inline void cover(int k, ll x, int l, int r) {
p[k].tag = x, p[k].sum = 1ll * (r - l + 1) * x;
}
inline void pushdown(int k, int l, int r) {
int ls = p[k].ls, rs = p[k].rs, mid = (l + r) >> 1;
if (p[k].tag) {
if (ls) cover(ls, p[k].tag, l, mid);
if (rs) cover(rs, p[k].tag, mid + 1, r);
}
}
void build(int &k, int l, int r) {
if (!k) {k = ++ cnt;}
if (l == r) {p[k].tag = p[k].sum = a[l]; return;}
int mid = (l + r) >> 1;
build(p[k].ls, l, mid); build(p[k].rs, mid + 1, r);
pushup(k);
}
void recover(int k, int l, int r, int x, int y, ll c) {
if (x <= l && r <= y) {cover(k, c, l, r); return;}
int mid = l + r >> 1; pushdown(k, l, r);
if (x <= mid) recover(p[k].ls, l, mid, x, y, c);
if (mid < y) recover(p[k].rs, mid + 1, r, x, y, c);
p[k].tag = 0; pushup(k);
}
void Change_to_Phi(int k, int l, int r, int x, int y) {
if (x <= l && r <= y && p[k].tag) {
cover(k, phi[p[k].tag], l, r); return;
}
int mid = (l + r) >> 1; pushdown(k, l, r);
if (x <= mid) Change_to_Phi(p[k].ls, l, mid, x, y);
if (mid < y) Change_to_Phi(p[k].rs, mid + 1, r, x, y);
pushup(k);
}
ll query(int k, int l, int r, int x, int y) {
if (x <= l && r <= y) return p[k].sum;
int mid = (l + r) >> 1; ll res = 0; pushdown(k, l, r);
if (x <= mid) res += query(p[k].ls, l, mid, x, y);
if (mid < y) res += query(p[k].rs, mid + 1, r, x, y);
return res;
}
int main() {
Get_Phi(1e7); scanf("%d", &T);
while (T --) {
scanf("%d%d", &n, &m);
for (int i = 1; i <= n; ++ i)
scanf("%d", &a[i]);
build(rt, 1, n); int id, l, r, x;
while (m --) {
scanf("%d%d%d", &id, &l, &r);
if (id == 1) Change_to_Phi(rt, 1, n, l, r);
else if (id == 2){
scanf("%d", &x); recover(rt, 1, n, l, r, x);
} else printf("%lld\n", query(rt, 1, n, l, r));
}
}
return 0;
}