模拟赛小结:2019-2020 ICPC, Asia Jakarta Regional Contest
比赛链接:传送门
离金最近的一次?,lh大佬carry场。
Problem A. Copying Homework 00:17(+) Solved by Dancepted
签到,读题有点慢了。而且配置vscode花了点时间。
#include <bits/stdc++.h>
using namespace std;
int a[100005];
int main() {
int n;
cin >> n;
for (int i = 1; i <= n; i++) {
scanf("%d", &a[i]);
}
for (int i = 1; i <= n; i++) {
printf("%d%c", n + 1 - a[i], " \n"[i==n]);
}
return 0;
}
Problem C. Even Path 00:44(+) Solved by Dancepted
每个格子通过even path能到达的区域是一个矩形,预处理这个矩形的边界就ok。签到。
“NO”写成了“No”,WA1不算罚时开心死了。
#include <bits/stdc++.h>
#define N 100005
using namespace std;
int R[N], C[N];
int l[N], r[N], u[N], d[N];
int main() {
int n, q;
cin >> n >> q;
for (int i = 1; i <= n; i++)
scanf("%d", &R[i]);
for (int i = 1; i <= n; i++)
scanf("%d", &C[i]);
l[1] = u[1] = 1;
for (int i = 2; i <= n; i++) {
if (R[i]%2 == R[i-1]%2)
u[i] = u[i - 1];
else
u[i] = i;
if (C[i]%2 == C[i-1]%2)
l[i] = l[i - 1];
else
l[i] = i;
}
r[n] = d[n] = n;
for (int i = n - 1; i >= 1; i--) {
if (R[i]%2 == R[i+1]%2)
d[i] = d[i + 1];
else
d[i] = i;
if (C[i]%2 == C[i+1]%2)
r[i] = r[i + 1];
else
r[i] = i;
}
while (q--) {
int ra, ca, rb, cb;
scanf("%d%d%d%d", &ra, &ca, &rb, &cb);
if (u[ra] <= rb && rb <= d[ra] && l[ca] <= cb && cb <= r[ca])
puts("YES");
else
puts("NO");
}
return 0;
}
Problem H. Twin Buildings 01:03(-1) Solved by xk
xk说是sb题。。。qwq。
#include<bits/stdc++.h>
using namespace std;
#define forn(i, n) for(int i = 0; i < (n); i++)
#define forab(i, a, b) for(int i = (a); i <= (b); i++)
#define forba(i, b, a) for(int i = (b); i >= (a); i--)
#define mst(array, Num, Kind, Count) memset(array, Num, sizeof(Kind) * (Count))
#define mp(x, y) make_pair(x, y)
#define fi first
#define se second
typedef long long ll;
typedef double db;
typedef pair<int, int> pii;
const int maxn = 1e5 + 5;
pii a[maxn];
int mxh[maxn];
int main()
{
int n;
ios::sync_with_stdio(0), cin.tie(0);
cin >> n;
ll ans = 0;
forn(i, n)
{
cin >> a[i].fi >> a[i].se;
if(a[i].fi > a[i].se)
swap(a[i].fi, a[i].se);
ans = max(ans, (ll)a[i].fi * a[i].se);
}
sort(a, a + n);
forba(i, n - 1, 1)
{
mxh[i - 1] = max(mxh[i], a[i].se);
ans = max(ans, (ll)a[i - 1].fi * min(a[i - 1].se, mxh[i - 1]) * 2);
}
cout << ans / 2 << (ans % 2 ? ".5" : ".0") << endl;
}
Problem K. Addition Robot 01:28(+) Solved by lh
lh一拍脑袋说这个tm好像可以区间合并,那不是线段树随便搞?于是K就被秒了。
具体地:
若左子树对应区间的A和B是:
$\begin{bmatrix}A_{l}&B_{l}\end{bmatrix} = \begin{bmatrix}A&B\end{bmatrix} * \begin{bmatrix}f_{AtoAl} & f_{AtoBl}\\ f_{BtoAl} & f_{BtoBl}\end{bmatrix}$
而右子树对应区间的A和B是:
$\begin{bmatrix}A_{r}&B_{r}\end{bmatrix} = \begin{bmatrix}A&B\end{bmatrix} * \begin{bmatrix}f_{AtoAr} & f_{AtoBr}\\ f_{BtoAr} & f_{BtoBr}\end{bmatrix}$
则合并后的区间对应的A和B是:
$\begin{bmatrix}A_{o}&B_{o}\end{bmatrix} = \begin{bmatrix}A&B\end{bmatrix} * \begin{bmatrix}f_{AtoAl} & f_{AtoBl}\\ f_{BtoAl} & f_{BtoBl}\end{bmatrix} * \begin{bmatrix}f_{AtoAr} & f_{AtoBr}\\ f_{BtoAr} & f_{BtoBr}\end{bmatrix}$
线段树每个节点维护4个系数。对于交换一个区间内的A和B,就是把系数AtoA和BtoA交换,AtoB和BtoB交换。合并的时候两个子区间的系数矩阵乘一下就行。
#include<bits/stdc++.h>
using namespace std;
#define INF 0x3f3f3f3f
#define LLINF 0x3f3f3f3f3f3f3f3f
#define lowbit(x) ((-x) & x)
#define ffor(i, d, u) for (int i = (d); i <= (u); ++i)
#define _ffor(i, u, d) for (int i = (u); i >= (d); --i)
#define mst(array, Num, Kind, Count) memset(array, Num, sizeof(Kind) * (Count))
#define mp(x, y) make_pair(x, y)
#define fi first
#define se second
#define N 100005
#define M 1000005
typedef long long ll;
typedef double db;
typedef pair<ll, ll> pll;
typedef pair<int, int> pii;
typedef pair<db, db> pdd;
const db PI = acos(-1);
const ll MO = 1e9 + 7;
const ll Inv2 = (MO + 1) / 2;
const bool debug = true;
template <typename T>
inline void read(T &x)
{
x=0;char c;T t=1;while(((c=getchar())<'0'||c>'9')&&c!='-');
if(c=='-'){t=-1;c=getchar();}do(x*=10)+=(c-'0');while((c=getchar())>='0'&&c<='9');x*=t;
}
template <typename T, typename... Args>
inline void read(T &x, Args &... args)
{
read(x), read(args...);
}
template <typename T>
inline void write(T x)
{
int len=0;char c[21];if(x<0)putchar('-'),x*=(-1);
do{++len;c[len]=(x%10)+'0';}while(x/=10);_ffor(i,len,1)putchar(c[i]);
}
int n, q;
char s[N];
struct node
{
ll a, b, x, y;
bool lazy;
node operator * (const node &other)
{
node ans;
ans.lazy = false;
ans.a = (a * other.a % MO + x * other.b % MO) % MO;
ans.b = (b * other.a % MO + y * other.b % MO) % MO;
ans.x = (a * other.x % MO + x * other.y % MO) % MO;
ans.y = (b * other.x % MO + y * other.y % MO) % MO;
return ans;
}
} t[N << 2];
inline void pushup(int o)
{
int lo = o << 1, ro = o << 1 | 1;
t[o] = t[lo] * t[ro];
}
void build(int o = 1, int l = 1, int r = n)
{
t[o].lazy = false;
if (l == r)
{
if (s[l] == 'A')
t[o].a = 1, t[o].b = 1, t[o].x = 0, t[o].y = 1;
else
t[o].a = 1, t[o].b = 0, t[o].x = 1, t[o].y = 1;
return;
}
int mid = (l + r) >> 1;
build(o << 1, l, mid), build(o << 1 | 1, mid + 1, r);
pushup(o);
}
inline void modify(int o)
{
t[o].lazy = !t[o].lazy;
swap(t[o].a, t[o].b), swap(t[o].x, t[o].y);
swap(t[o].a, t[o].x), swap(t[o].b, t[o].y);
}
inline void pushdown(int o)
{
if (t[o].lazy == false)
return;
int lo = o << 1, ro = o << 1 | 1;
modify(lo), modify(ro), t[o].lazy = false;
}
void change(int cl, int cr, int o = 1, int l = 1, int r = n)
{
if (cl == l && cr == r)
{
modify(o);
return;
}
pushdown(o);
int mid = (l + r) >> 1;
if (mid >= cr)
change(cl, cr, o << 1, l, mid);
else if (mid < cl)
change(cl, cr, o << 1 | 1, mid + 1, r);
else
change(cl, mid, o << 1, l, mid), change(mid + 1, cr, o << 1 | 1, mid + 1, r);
pushup(o);
}
node query(int ql, int qr, int o = 1, int l = 1, int r = n)
{
if (ql == l && qr == r)
return t[o];
pushdown(o);
int mid = (l + r) >> 1;
if (mid >= qr)
return query(ql, qr, o << 1, l, mid);
else if (mid < ql)
return query(ql, qr, o << 1 | 1, mid + 1, r);
else
return query(ql, mid, o << 1, l, mid) * query(mid + 1, qr, o << 1 | 1, mid + 1, r);
}
inline int ac()
{
read(n, q);
scanf("%s", s + 1);
build();
while (q--)
{
int op, l, r, a, b;
read(op, l, r);
if (op == 2)
{
read(a, b);
node ans = query(l, r);
write((a * ans.a % MO + b * ans.b % MO) % MO), putchar(' '), write((a * ans.x % MO + b * ans.y % MO) % MO), putchar('\n');
}
else
change(l, r);
}
return 0;
}
int main()
{
ac();
return 0;
}
Problem G. Performance Review 02:44(+) Solved by lh
据说是个线段树之类的题,lh分分钟就秒了,我和xk站在一边人都看傻了。
#include<bits/stdc++.h>
using namespace std;
#define INF 0x3f3f3f3f
#define lowbit(x) ((-x) & x)
#define ffor(i, d, u) for (int i = (d); i <= (u); ++i)
#define _ffor(i, u, d) for (int i = (u); i >= (d); --i)
#define mst(array, Num, Kind, Count) memset(array, Num, sizeof(Kind) * (Count))
#define mp(x, y) make_pair(x, y)
#define fi first
#define se second
#define N 100005
#define M 3000005
typedef long long ll;
typedef double db;
typedef pair<ll, ll> pll;
typedef pair<int, int> pii;
typedef pair<db, db> pdd;
const db PI = acos(-1);
const ll MO = 1e9 + 7;
const ll Inv2 = (MO + 1) / 2;
const bool debug = true;
template <typename T>
inline void read(T &x)
{
x=0;char c;T t=1;while(((c=getchar())<'0'||c>'9')&&c!='-');
if(c=='-'){t=-1;c=getchar();}do(x*=10)+=(c-'0');while((c=getchar())>='0'&&c<='9');x*=t;
}
template <typename T, typename... Args>
inline void read(T &x, Args &... args)
{
read(x), read(args...);
}
template <typename T>
inline void write(T x)
{
int len=0;char c[21];if(x<0)putchar('-'),x*=(-1);
do{++len;c[len]=(x%10)+'0';}while(x/=10);_ffor(i,len,1)putchar(c[i]);
}
int n, m, q, a = 0;
vector<int> b[N];
int minx[N << 2], lazy[N << 2];
void build(int o = 1, int l = 1, int r = m)
{
minx[o] = a, lazy[o] = 0;
if (l == r)
return;
int mid = (l + r) >> 1;
build(o << 1, l, mid), build(o << 1 | 1, mid + 1, r);
}
inline void pushdown(int o)
{
if (lazy[o] == 0)
return;
int lo = o << 1, ro = o << 1 | 1;
minx[lo] += lazy[o], minx[ro] += lazy[o];
lazy[lo] += lazy[o], lazy[ro] += lazy[o];
lazy[o] = 0;
}
void change(int ql, int qr, int x, int o = 1, int l = 1, int r = m)
{
if (l == ql && r == qr)
{
lazy[o] += x, minx[o] += x;
return;
}
pushdown(o);
int mid = (l + r) >> 1;
if (mid >= qr)
change(ql, qr, x, o << 1, l, mid);
else if (mid < ql)
change(ql, qr, x, o << 1 | 1, mid + 1, r);
else
change(ql, mid, x, o << 1, l, mid), change(mid + 1, qr, x, o << 1 | 1, mid + 1, r);
minx[o] = min(minx[o << 1], minx[o << 1 | 1]);
}
inline int ac()
{
int r, x, y, z;
read(n, m, q, x);
ffor(i, 2, n) read(y), a += (y < x ? 1 : 0);
build(), a = 0;
ffor(i, 1, m)
{
read(r);
if (a)
change(i, m, a);
a = 0;
ffor(j, 1, r)
{
read(y), b[i].emplace_back(y);
a += (y < x ? 1 : 0);
}
change(i, m, -r);
}
int mu = x;
while (q--)
{
read(x, y, z);
if ((b[x][y - 1] > mu) == (z > mu) || x == m)
{
b[x][y - 1] = z;
puts(minx[1] >= 0 ? "1" : "0");
continue;
}
if (z > mu)
change(x + 1, m, -1);
else
change(x + 1, m, 1);
b[x][y - 1] = z;
puts(minx[1] >= 0 ? "1" : "0");
}
return 0;
}
int main()
{
ac();
return 0;
}
后面我和xk去开了E,我想了个贪心,但是有些细节没理清楚,而xk怎么看怎么像差分约束,对着板子敲了一下,但是后面发现不太会改板子的样子。后面还尝试写F和L。
F我看了一眼割出来的点必须是树的重心,但是不知道怎么判子树同构,瞎猜了一个做法,但是bug有点多没来得及调完。(赛后调完交上去WA17,应该是个假做法,明早补)
L大概是个图论,但是xk最后没来得及写。
总结:
这场其实有点不太认真,因为各种原因过了10分钟左右才开始看题。中间我还接到了移动的电话骗我去改套餐(结果是兼职的同学找人代打电话,细节没交代好,害我白跑一趟),还顺路买了个晚饭。。。
我后面上厕所也上了挺久的,浪费了不少时间。
最后手上还有三道疑似可做的题,时间多一点说不定能再过一题,这周六要认真点打了QAQ。
还有个人感觉xk看E的时候的:“这肯定是个差分约束,但是我不会差分约束”,这种做法不太好,我感觉这样很容易看成假算法。我有一场CF就是死脑筋,那场的D我就说:“这肯定是个马拉车,但是我不会马拉车,我去看一下我的板子”,但是实际上D是个思维题,我还因此掉了不少分。
不过xk和我切水题又快起来了的样子qwq。然后lh一眼一道线段树,分分钟过掉真的nb没话说。
赛后想想,F题的判树的同构的算法是猜出来的,而xk的L应该是证明过的比较严谨,下次碰到的时候应该优先让这种题先敲。
其他的。。。和某位肾结石的好基友讨论了一下,才发现我们俩从初中开始就熬夜成瘾了,身体真的吃不消,以后还是早睡早起吧QWQ。