/*  Hi,I am warabi. Welcome to my coding space. Let me give you a great word of advice. "The pen is mightier than the sword" -by Edward Bulwer-Lytton   _,,....,,_  _人人人人人人人人人人人人人人人人人人_ -''":::::::::::::`''>   ゆっくりできるとでも?  < ヽ::::::::::::::::::::: ̄^Y^Y^Y^Y^Y^Y^Y^Y^Y^Y ̄  |::::::;ノ´ ̄\:::::::::::\_,. -‐ァ     __   _____   ______  |::::ノ   ヽ、ヽr-r'"´  (.__    ,´ _,, '-´ ̄ ̄`-ゝ 、_ イ、 _,.!イ_  _,.ヘーァ'二ハ二ヽ、へ,_7   'r ´          ヽ、ン、 ::::::rー''7コ-‐'"´    ;  ', `ヽ/`7 ,'==─-      -─==', i r-'ァ'"´/  /! ハ  ハ  !  iヾ_ノ i イ iゝ、イ人レ/_ルヽイ i | !イ´ ,' | /__,.!/ V 、!__ハ  ,' ,ゝ レリイi (ヒ_]     ヒ_ン ).| .|、i .|| `!  !/レi' (ヒ_]     ヒ_ン レ'i ノ   !Y!""  ,___,   "" 「 !ノ i | ,'  ノ   !'"  ,___,  "' i .レ'    L.',. ヽ _ン    L」 ノ| .|  (  ,ハ    ヽ _ン   人!      | ||ヽ、       ,イ| ||イ| / ,.ヘ,)、  )>,、 _____, ,.イ  ハ    レ ル` ー--─ ´ルレ レ´  Take your time.    GIVE ME AC!!!!!!!!!!!!!!!!!!!!!!!!!   */ #include using namespace std; #define ll long long #define ld long double #define V vector #define rep(i,a,b) for(ll i=a;i=b;i--) #define reph(i,vec) for(auto &i:vec) #define FI first #define SE second #define P pair #define PB push_back #define EB emplace_back #define INS insert #define all(vec) vec.begin(),vec.end() #define in(name) ll name;cin >> name #define ins(name) string name;cin >> name; #define inc(name) char name;cin >> name #define ing(name,size,E) vector> name(size);rep(i,0,E){in(a);in(b);a--;b--;name[a].PB(b);name[b].PB(a);} #define ing_on(name,size,E) vector> name(size);rep(i,0,E){in(a);in(b);a--;b--;name[a].PB(b);} #define ing_cost(name,size,E) vector>> name(size);rep(i,0,E){in(a);in(b);in(c);a--;b--;name[a].PB({b,c});name[b].PB({a,c});} #define invll(name,size) vector name(size);reph(i,name)cin >> i #define invvll(name,size1,size2) vector> name(size1,vector(size2));reph(v,name)reph(i,v)cin>>i; #define invp(name,size) vector> name(size);for(ll i=0;i> name[i].FI >> name[i].SE #define out(n) cout << (n) << endl #define _out(n) cout << " " << (n) << endl; #define notout(n) cout << (n) #define out_(n) cout << (n) << " " #define set_out(n,k) cout << fixed << setprecision(k) << (n) << endl; #define set_notout(n,k) cout << fixed << setprecision(k) << (n) << endl; #define set_out_(n,k) cout << fixed << setprecision(k) << (n) << " "; #define vout(vec) for(int i=0;i0) #define INF 1000000000000000000ll #define MOD 1000000007 #define MOD2 998244353 #define CMOD MOD2 #define EPS 0.00000001 //debug #define bug assert(false); #define debug cout<<"OK Line "<<__LINE__< inline bool chmax(T& a, const T b) { if (a < b) { a = b; return true; }return false; } template inline bool chmin(T& a, const T b) { if (a > b) { a = b; return true; }return false; } // namespace Warabi { //常備変数のコーナー V primes(1e7+5, true); V prime_list; V prime_rui(1e7+5, 0LL); V visiteded(300100); V afed(300100, false); V k1(200100, 0ll); V k2(200100, 0ll); // //常備構造体宣言のコーナー class UnionFind { private: ll NumberOfElements; ll Not1NumberOfelements; public: vector par; vector SIZE; void init(ll sz) { par.resize(sz, -1LL); SIZE.resize(sz, 1LL); NumberOfElements = sz; Not1NumberOfelements = 0ll; } ll root(ll pos) { if (par[pos] == -1) return pos; par[pos] = root(par[pos]); return par[pos]; } bool same(ll u, ll v) { if (root(u) == root(v)) return true; return false; } ll SZ(ll u) { return SIZE[root(u)]; } ll noe() { return NumberOfElements; } ll nnoe() { return Not1NumberOfelements; } bool unite(ll u, ll v) { u = root(u); v = root(v); if (u == v) { return false; } if (SZ(u) == 1LL and SZ(v) == 1LL)Not1NumberOfelements++; if (u < v)swap(u, v); par[u] = v; SIZE[v] += SIZE[u]; NumberOfElements--; return true; } }; class SCC { public: V> G; V> UG; V order; V GROUP; V visited; ll sz_count; V groupsize; void init(ll sz) { G.resize(sz, V(0)); UG.resize(sz, V(0)); GROUP.resize(sz, -1ll); visited.resize(sz, false); sz_count = 0LL; return; } void dfs(ll now) { visited[now] = true; reph(i, G[now]) { if (visited[i])continue; dfs(i); } order.PB(now); return; } void dfs2(ll now, ll group) { GROUP[now] = group; sz_count++; reph(i, UG[now]) { if (GROUP[i] != -1ll and GROUP[i] != group)continue; if (GROUP[i] != -1ll)continue; dfs2(i, group); } return; } void make_group(V> Graph_function) { G = Graph_function; rep(i, 0, (ll)G.size()) { reph(j, G[i]) { UG[j].PB(i); } } rep(i, 0, (ll)G.size()) { if (visited[i])continue; dfs(i); } reverse(all(order)); ll now_group = 0LL; reph(i, order) { if (GROUP[i] != -1)continue; ll prev = sz_count; dfs2(i, now_group); now_group++; groupsize.PB(sz_count - prev); } return; } }; template class SegmentTree { public: long long sz; using FX = function; using FA = function; using FM = function; const FX fx; const FA fa; const FM fm; const X ex; const M em; vector node; vector lazy; SegmentTree(long long sz_, FX fx_, FA fa_, FM fm_, X ex_, M em_) :sz(), fx(fx_), fa(fa_), fm(fm_), ex(ex_), em(em_) { long long n = 1LL; while (n < sz_)n *= 2; sz = n; node.resize(sz * 2 - 1, ex); lazy.resize(sz * 2 - 1, em); } void set(long long index, X x) { node[index + sz - 1] = x; return; } void build() { for (int i = sz - 2; i >= 0; i--)node[i] = fx(node[i * 2 + 1], node[i * 2 + 2]); return; } void eval(long long k) { if (lazy[k] == em)return; if (k < sz - 1) { lazy[k * 2 + 1] = fm(lazy[k * 2 + 1], lazy[k]); lazy[k * 2 + 2] = fm(lazy[k * 2 + 2], lazy[k]); } node[k] = fa(node[k], lazy[k]); lazy[k] = em; } void update_sub(long long a, long long b, M x, long long k, long long l, long long r) { //cout << " " << a << " " << b << " " << l << " " << r << endl; eval(k); if (a <= l and r <= b) { lazy[k] = fm(lazy[k], x); //cout<<" "< class multimap { private: map mp; public: void add(ll x) { mp[x]++; return; } void del(ll x) { mp[x]--; if (mp[x] < 1)mp.erase(x); return; } T maximum() { return mp.rbegin()->first; } T minimum() { return mp.begin()->first; } }; class LCA { public: vector> parent; vector dist; vector> G; LCA(const vector>& gr) { init(gr); } void dfs(long long v, long long p, long long d) { parent[0][v] = p; dist[v] = d; for (long long next : G[v]) { if (next == p)continue; dfs(next, v, d + 1); } return; } void init(const vector>& gr) { G = gr; parent.assign(33, vector(G.size(), -1LL)); dist.assign(G.size(), -1LL); dfs(0LL, -1LL, 0LL); for (int i = 0; i < 32; i++) { for (int j = 0; j < (int)G.size(); j++) { if (parent[i][j] < 0LL) { parent[i + 1][j] = -1LL; } else { parent[i + 1][j] = parent[i][parent[i][j]]; } } } return; } long long lca(long long u, long long v) { if (dist[u] < dist[v])swap(u, v); long long between = dist[u] - dist[v]; for (int i = 0; i < 33; i++) { if (between & (1LL << i)) { u = parent[i][u]; } } if (u == v)return u; for (int i = 32; i >= 0; i--) { if (parent[i][u] != parent[i][v]) { u = parent[i][u]; v = parent[i][v]; } } assert(parent[0][u] == parent[0][v]); return parent[0][u]; } long long get_dist(long long u, long long v) { return dist[u] + dist[v] - 2 * dist[lca(u, v)]; } bool on_path(long long u, long long v, long long a) { return get_dist(u, v) == get_dist(u, a) + get_dist(v, a); } }; class nCk { public: const long long m; const long long MAXIMUM = 2000005; vector fac, facinv, inv; nCk(long long m_); ~nCk(); long long com(long long n, long long k)const; }; nCk::nCk(long long m_) :m(m_) { fac.resize(MAXIMUM + 3); facinv.resize(MAXIMUM + 3); inv.resize(MAXIMUM + 3); fac[0] = fac[1] = 1; facinv[0] = facinv[1] = 1; inv[1] = 1; for (long long i = 2; i < MAXIMUM + 2; i++) { fac[i] = fac[i - 1] * i % m; inv[i] = m - inv[m % i] * (m / i) % m; facinv[i] = facinv[i - 1] * inv[i] % m; } } nCk::~nCk() {} long long nCk::com(long long n, long long k)const { if (k == 0)return 1; if (n < k)return 0LL; assert(!(n < 0 || k < 0)); return fac[n] * (facinv[k] * facinv[n - k] % m) % m; } // //常備構造体宣言のコーナー // //常備関数のコーナー void Yes(bool f) { cout << (f ? "Yes" : "No") << endl; } void YES(bool f) { cout << (f ? "YES" : "NO") << endl; } //木の直径を求める tuple Cdiameter(V>> G, ll start, bool type) { visiteded = afed; queue sirabe; sirabe.push(start); V short_load(G.size(), 0LL); while (!sirabe.empty()) { ll s = sirabe.front(); sirabe.pop(); visiteded[s] = true; reph(i, G[s]) { if (visiteded[i.FI])continue; short_load[i.FI] = short_load[s] + i.SE; sirabe.push(i.FI); } } ll far_max = -1LL; ll element = -1LL; rep(i, 0, (ll)G.size()) { if (far_max >= short_load[i])continue; far_max = short_load[i]; element = i; } if (type)return Cdiameter(G, element, false); else return { start,element,far_max }; } //素数の取得 void prime_init() { const int Limit=1e7+5; V at(Limit, true); primes = at; primes[0] = primes[1] = false; rep(i, 2, Limit) { if (!primes[i])continue; for (ll j = i * 2; j <= Limit; j += i)primes[j] = false; } rep(i, 1, Limit) { if (primes[i]) { prime_list.PB(i); prime_rui[i] = prime_rui[i - 1] + 1; } else { prime_rui[i] = prime_rui[i - 1]; } } return; } //modpow long long modpow(long long a, long long b, long long m) { long long p = 1, q = a; for (int i = 0; i < 63; i++) { if ((b & (1LL << i)) != 0) { p *= q; p %= m; } q *= q; q %= m; } return p; } //逆元 long long Div(long long a, long long b, long long m) { return (a * modpow(b, m - 2, m)) % m; } //点と点の距離を返す long double Dis(ll ax, ll ay, ll bx, ll by) { return sqrt(pow(ax - bx, 2) + pow(ay - by, 2)); } //二つのベクトルから平行四辺形の面積を返す ll he(ll x0, ll y0, ll x1, ll y1, ll x2, ll y2) {//外積を二で割る return abs((x1 - x0) * (y2 - y0) - (x2 - x0) * (y1 - y0)); } // template ll Lis_size(V arr, ll sz) { const T TINF = numeric_limits::max(); V DP(sz + 1, TINF); DP[0] = -TINF; rep(i, 0, sz) { *lower_bound(all(DP), arr[i]) = arr[i]; } ll ans = 0LL; rep(i, 1, sz + 1) { if (DP[i] != TINF)ans = i; } return ans; } string toBinary(ll n) { string r; while (n != 0LL) { r += (n % 2LL == 0LL ? "0" : "1"); n /= 2LL; } return r; } template V press(V arr) { V X = arr; sort(all(X)); X.erase(unique(all(X)), X.end()); V ans(arr.size()); rep(i, 0LL, (ll)arr.size()) { ans[i] = lower_bound(all(X), arr[i]) - X.begin(); } return ans; } P>, V>>> warshall_floyd(ll N, V>> G) { V> DP(N, V(N, INF)); rep(i, 0, N)DP[i][i] = 0LL; V>> prev(N, V>(N, { INF,INF })); rep(i, 0, N) { reph(j, G[i]) { DP[i][j.FI] = j.SE; prev[i][j.FI] = { i,j.FI }; } } rep(k, 0, N) { rep(i, 0, N) { rep(j, 0, N) { if (chmin(DP[i][j], DP[i][k] + DP[k][j])) { prev[i][j] = prev[k][j]; } } } } return { DP,prev }; } template void to_sum(V& arr) { rep(i, 0, (ll)arr.size() - 1LL) { arr[i + 1] += arr[i]; } return; } template void including_map(ll H, ll W, V>& G, T c) { V> new_G(H + 2, V(W + 2)); rep(i, 0, W + 2)new_G[0][i] = c; rep(i, 1, H + 1) { new_G[i][0] = c; new_G[i][W + 1] = c; rep(j, 1, W + 1) { new_G[i][j] = G[i - 1][j - 1]; } } rep(i, 0, W + 2)new_G[H + 1][i] = c; G = new_G; return; } template class BIT { private: int n; vector bit; public: // 0_indexed で i 番目の要素に x を加える void add(int i, T x){ i++; while(i < n){ bit[i] += x, i += i & -i; } } // 0_indexed で [0,i] の要素の和(両閉区間!!) T sum(int i){ i++; T s = 0; while(i > 0){ s += bit[i], i -= i & -i; } return s; } BIT(){} //初期値がすべて0の場合 BIT(int sz) : n(sz+1), bit(n, 0){} BIT(const vector& v) : n((int)v.size()+1), bit(n, 0){ for(int i = 0; i < n-1; i++){ add(i,v[i]); } } void print(){ for(int i = 0; i < n-1; i++){ cout << sum(i) - sum(i-1) << " "; } cout << "\n"; } //-1スタート void print_sum(){ for(int i = 0; i < n; i++){ cout << sum(i-1) << " "; } cout << "\n"; } }; // u を昇順にソートするのに必要な交換回数(転倒数) (u は {0,..., n-1} からなる重複を許した長さ n の数列) long long inv_count(const vector& u) { int n = (int)u.size(); BIT bt(n); long long ans = 0; for(int i = 0; i < n; i++){ ans += i - bt.sum(u[i]); bt.add(u[i], 1); } return ans; } set factor(ll N) { set ans; for (ll i = 1LL; i * i <= N; i++) { if (N % i)continue; ans.INS(i); ans.INS(N / i); } return ans; } V dijkstra(ll sz, V>> G, ll s) { V ans(sz, INF); ans[s] = 0LL; priority_queue, vector>, greater>> examine; examine.push({ 0LL,s }); while (examine.size()) { P p = examine.top(); examine.pop(); ll now = p.SE, dist = p.FI; if (ans[now] < dist)continue; reph(i, G[now]) { ll next = i.FI, c = i.SE; if (chmin(ans[next], dist + c)) { examine.push({ dist + c,next }); } } } return ans; } V pass(const V>& G, ll s, ll t) { V ans, res; function dfs = [&](ll now,ll p) { res.PB(now); if (now == t) { ans = res; } reph(next, G[now]) { if (next==p)continue; dfs(next,now); } res.pop_back(); return; }; dfs(s,-1); return ans; } // 負の数にも対応した mod (a = -11 とかでも OK) inline long long mod(long long a, long long m) { long long res = a % m; if (res < 0) res += m; return res; } // 拡張 Euclid の互除法 long long extGCD(long long a, long long b, long long& p, long long& q) { if (b == 0) { p = 1; q = 0; return a; } long long d = extGCD(b, a % b, q, p); q -= a / b * p; return d; } long long extGcd(long long a, long long b, long long& p, long long& q) { if (b == 0) { p = 1; q = 0; return a; } long long d = extGcd(b, a % b, q, p); q -= a / b * p; return d; } // 逆元計算 (ここでは a と m が互いに素であることが必要) long long modinv(long long a, long long m) { long long x, y; extGCD(a, m, x, y); return mod(x, m); // 気持ち的には x % m だが、x が負かもしれないので } // Garner のアルゴリズム, x%MOD, LCM%MOD を求める (m は互いに素でなければならない) // for each step, we solve "coeffs[k] * t[k] + constants[k] = b[k] (mod. m[k])" // coeffs[k] = m[0]m[1]...m[k-1] // constants[k] = t[0] + t[1]m[0] + ... + t[k-1]m[0]m[1]...m[k-2] long long Garner(vector b, vector m, long long Mmod) { m.push_back(Mmod); // banpei vector coeffs((int)m.size(), 1); vector constants((int)m.size(), 0); for (int k = 0; k < (int)b.size(); ++k) { long long t = mod((b[k] - constants[k]) * modinv(coeffs[k], m[k]), m[k]); for (int i = k + 1; i < (int)m.size(); ++i) { (constants[i] += t * coeffs[i]) %= m[i]; (coeffs[i] *= m[k]) %= m[i]; } } return constants.back(); } pair ChineseRem(const vector& b, const vector& m) { long long r = 0, M = 1; for (int i = 0; i < (int)b.size(); ++i) { long long p, q; long long d = extGcd(M, m[i], p, q); // p is inv of M/d (mod. m[i]/d) if ((b[i] - r) % d != 0) return make_pair(0, -1); long long tmp = (b[i] - r) / d * p % (m[i] / d); r += M * tmp; M *= m[i] / d; } return make_pair(mod(r, M), M); } map bunkai(ll n) { map ans; ll Nn = n; for (int i = 2; i * i <= Nn; i++) { while (n % i == 0) { n /= i; ans[i]++; } } if (n > 1) { ans[n]++; } return ans; } template void RLE(vector& arr) { vector res; res.push_back(arr[0]); for (const T t : arr) { if (res.back() != t)res.push_back(t); } arr = res; return; } vector> EulerTour(const vector>& G) { long long N = (long long)G.size(); vector> res(N); long long now = 0ll; function dfs = [&](long long v, long long p) { res[v].first = now; ++now; for (const long long next : G[v]) { if (next == p)continue; dfs(next, v); } res[v].second = now; ++now; return; }; dfs(0, -1ll); return res; } template struct BIT2D { int H, W; vector> bit; // データの格納先 BIT2D(int H_, int W_) { init(H_, W_); } void init(int H_, int W_) { H = H_ + 1; W = W_ + 1; bit.assign(H, vector(W, 0)); } void add(int h, int w, T x) { for (int i = h; i < H; i += (i & -i)) { for (int j = w; j < W; j += (j & -j)) { bit[i][j] += x; } } } // 1≦i≦h かつ 1≦j≦w T sum(int h, int w) { T s(0); for (int i = h; i > 0; i -= (i & -i)) { for (int j = w; j > 0; j -= (j & -j)) { s += bit[i][j]; } } return s; } // h1≦i

left, right; const int width; void add(const int id); void del(const int id); public: vector ans; Mo(const int n) : width((int)sqrt(n)) {} //クエリ[l,r) void insert(const int l, const int r) { left.push_back(l), right.push_back(r); } void solve() { const int sz = (int)left.size(); int nl = 0, nr = 0; vector ord(sz); iota(ord.begin(), ord.end(), 0); sort(ord.begin(), ord.end(), [&](const int a, const int b) { const int c = left[a] / width, d = left[b] / width; return (c == d) ? ((c & 1) ? (right[b] < right[a]) : (right[a] < right[b])) : (c < d); }); ans.resize(sz); for (const int id : ord) { while (nl > left[id]) add(--nl); while (nr < right[id]) add(nr++); while (nl < left[id]) del(nl++); while (nr > right[id]) del(--nr); ans[id] = res; } } }; //idは元の配列のインデックス void Mo::add(const int id) { res -= cnt[a[id]] / 2; res += (++cnt[a[id]]) / 2; } void Mo::del(const int id) { res -= cnt[a[id]] / 2; res += (--cnt[a[id]]) / 2; } constexpr long long mod=MOD; class mint { long long x; public: mint(long long x=0) : x((x%mod+mod)%mod) {} mint operator-() const { return mint(-x); } mint& operator+=(const mint& a) { if ((x += a.x) >= mod) x -= mod; return *this; } mint& operator-=(const mint& a) { if ((x += mod-a.x) >= mod) x -= mod; return *this; } mint& operator*=(const mint& a) { (x *= a.x) %= mod; return *this; } mint operator+(const mint& a) const { mint res(*this); return res+=a; } mint operator-(const mint& a) const { mint res(*this); return res-=a; } mint operator*(const mint& a) const { mint res(*this); return res*=a; } mint pow(ll t) const { if (!t) return 1; mint a = pow(t>>1); a *= a; if (t&1) a *= *this; return a; } // for prime mod mint inv() const { return pow(mod-2); } mint& operator/=(const mint& a) { return (*this) *= a.inv(); } mint operator/(const mint& a) const { mint res(*this); return res/=a; } friend ostream& operator<<(ostream& os, const mint& m){ os << m.x; return os; } }; } // long long gcd(long long a, long long b) { if (b == 0LL)return a; return gcd(b, a % b); } long long lcm(long long a, long long b) { return a * b / gcd(a, b); } signed main() { /* 文章を読み落としていませんか? 制約も隅々まで読んでいますか? 注意 ・セグ木のupdate関数はl,rの値を渡したときにl以上r未満の区間だからご注意を  rは含みません ・BITって1-indexedなんだぜ  イケてるよな ・using namespace Warabi?? */ //using namespace Warabi; //mintのMODは確認した? in(N);in(M);in(L);--L; invll(T,N);ing_cost(G,N,M); V> dist(N);rep(i,0,N)dist[i]=Warabi::dijkstra(N,G,i); ll ans=INF; rep(to,0,N){ //集める頂点=to ll res=0ll; rep(i,0,N)res+=dist[to][i]*T[i]*2; ll extra=dist[to][L]; rep(i,0,N){ //toに向かう途中でトラックによる if(T[i])chmin(extra,dist[L][i]-dist[i][to]); } chmin(ans,res+extra); } out(ans); }