从终点逆推,d[u]表示进入u以后剩下的货物,那么进入u之前的货物数量设为y,d[u] = x,那么y-x=ceil(y/20.0)=(y-1)/20+1=(y+19)/20。
(y-x)*20+r=y+19,0≤r≤19,即19*y=20*x+r,根据题意y应该尽量小,x的部分是不能变动的,所以y=x+ceil(x/19.0)。
然后从起点找一条字典序最小的路径即可,因为每个字母都是独一无二的,所以不必bfs,每次记录一个点就够了。
#include<bits/stdc++.h> using namespace std; const int maxn = 54, maxm = maxn*maxn, town = 0, village = 1; bool g[maxn][maxn]; int head[maxn],nxt[maxm],to[maxm],ecnt; int tp[maxn]; int id[256]; char rid[256]; void addEdge(int u,int v) { to[ecnt] = v; nxt[ecnt] = head[u]; head[u] = ecnt++; } int id_cnt; void init() { memset(head,-1,sizeof(head)); memset(id,-1,sizeof(id)); memset(g,0,sizeof(g)); ecnt = 0; id_cnt = 0; } inline int ID(char c) { if(~id[c]) return id[c]; id[c]= id_cnt; rid[id_cnt] = c; tp[id_cnt] = 'a'<=c&&c<='z';//写成isalpha(c)-1;WA了 return id_cnt++; } typedef long long ll; typedef pair<ll,int> Node; #define fi first #define se second ll d[maxn]; void dijkstra(int s,ll p) { memset(d,0x7f,sizeof(d)); priority_queue<Node,vector<Node>,greater<Node> > q; q.push(Node(d[s] = p,s)); while(q.size()){ Node x = q.top(); q.pop(); int u = x.se; if(d[u] != x.fi) continue; ll t = (tp[u]?(1+d[u]):((d[u]+18)/19+d[u])); for(int i = head[u]; ~i; i = nxt[i]){ int v = to[i]; if(d[v] > t ){ q.push(Node(d[v]= t,v)); } } } } ll cost(int u,int v) { return tp[v]?1:((d[u]+19)/20); } void FindPath(int s,int e) { int u = s; while(u != e){ printf("%c-",rid[u]); int nex = -1; for(int i = head[u]; ~i; i = nxt[i]){ int v = to[i]; if(d[u]- cost(u,v) >= d[v]){ if(~nex) { if(rid[v] < rid[nex]) nex = v; }else nex = v; } } swap(nex,u); } printf("%c ",rid[e]); } int main() { //freopen("in.txt","r",stdin); int n; char a[10],b[10]; int kas = 0; while(scanf("%d",&n),~n){ init(); while(n--){ scanf("%s%s",a,b); int u = ID(*a), v = ID(*b); if(!g[u][v]){ g[u][v] = g[v][u] = true; addEdge(u,v); addEdge(v,u); } } ll p; scanf("%lld%s%s",&p,a,b); int s = ID(*a),e = ID(*b); dijkstra(e,p); printf("Case %d: %lld ",++kas,d[s]); FindPath(s,e); } return 0; }