标签:cost 时间复杂度 blog 最大 顶点 art 函数 als bfs
假设每条边除了有一个容量限制外,还有一个单位流量所需的费用(cost)。该网络中花费最小的最大流称为最小费用最大流,即总流量最大的情况下,总费用最小的流。
和 Edmonds-Karp 算法类似,但每次用 Bellman-Ford 算法而非 BFS 找增广路。只要初始流是该流量下的最小费用可行流,每次增广后的新流都是新流量下的最小费用流。
如何用优先队列代替Bellman-Ford中的普通队列就是SPFA版。
struct Edge{ int from, to, cap, flow, cost; Edge(int u, int v, int c, int f, int w):from(u), to(v), cap(c), flow(f), cost(w){} }; struct MCMF { int n, m; vector<Edge>edges; vector<ll>G[maxn]; int inq[maxn]; //是否在队列中 int d[maxn]; //Bellman-Ford int p[maxn]; //上一条弧 int a[maxn]; //可改进量 void init(int n) { this->n = n; for(int i = 0;i < n;i++) G[i].clear(); edges.clear(); } void AddEdge(int from, int to, int cap, int cost) { edges.push_back(Edge(from, to, cap, 0, cost)); //有向边 edges.push_back(Edge(to ,from, 0, 0, -cost)); m = edges.size(); G[from].push_back(m-2); G[to].push_back(m-1); } bool BellmanFord(int s, int t, int& flow, int& cost) { for(int i = 0;i < n;i++) d[i] = INF; memset(inq, 0, sizeof(inq)); d[s] = 0; inq[s] =1; p[s] = 0;a[s] = INF; queue<int>Q;; Q.push(s); while(!Q.empty()) { int u = Q.front(); Q.pop(); inq[u] = 0; for(int i = 0;i < G[u].size();i++) { Edge& e = edges[G[u][i]]; if(e.cap > e.flow && d[e.to] > d[u] + e.cost) { d[e.to] = d[u] + e.cost; p[e.to] = G[u][i]; a[e.to] = min(a[u], e.cap - e.flow); if(!inq[e.to]){ Q.push(e.to); inq[e.to] = 1;} } } } if(d[t] == INF) return false; flow += a[t]; cost += d[t] * a[t]; for(int u = t;u != s;u = edges[p[u]].from) { edges[p[u]].flow += a[t]; edges[p[u]^1].flow -=a[t]; } return true; } //需要保证初始网络中没有负权圈 int MaxcostMaxflow(int s, int t, int& cost) { int flow = 0; cost = 0; while(BellmanFord(s, t, flow, cost)); return flow; } }mcmf;
可以采用加入“势函数”后的Dijkstra算法。因为对于每一条边 $e=(u, v)$,有如下事实成立:$h(v) \leq h(u)+e.cost$ (其中 $h(u)$ 表示 $s$ 到 $u$ 的最短距离),因此令 $dis[v] = dis[u] + e.cost + h[u] - h[v]$,那么所有的 $dist$ 值必然大于等于0,这样就能用 $Dijkstra$ 求解了。
下面代码中用了一个优先队列,每次优先队列列出 $dist$ 值小的元素。
整个算法的时间复杂度为 $O(FElogV)$($F$ 是流量,$E$ 是边数,$V$ 是顶点数)。
struct edge { int to, capacity, cost, rev; edge() {} edge(int to, int _capacity, int _cost, int _rev) :to(to), capacity(_capacity), cost(_cost), rev(_rev) {} }; struct Min_Cost_Max_Flow { int V, H[maxn + 5], dis[maxn + 5], PreV[maxn + 5], PreE[maxn + 5]; vector<edge> G[maxn + 5]; //调用前初始化 void Init(int n) { V = n; for (int i = 0; i <= V; ++i)G[i].clear(); } //加边 void Add_Edge(int from, int to, int cap, int cost) { G[from].push_back(edge(to, cap, cost, G[to].size())); G[to].push_back(edge(from, 0, -cost, G[from].size() - 1)); } //flow是自己传进去的变量,就是最后的最大流,返回的是最小费用 int Min_cost_max_flow(int s, int t, int f, int& flow) { int res = 0; fill(H, H + 1 + V, 0); while (f) { priority_queue <pair<int, int>, vector<pair<int, int>>, greater<pair<int, int>> > q; fill(dis, dis + 1 + V, INF); dis[s] = 0; q.push(pair<int, int>(0, s)); while (!q.empty()) { pair<int, int> now = q.top(); q.pop(); int v = now.second; if (dis[v] < now.first)continue; for (int i = 0; i < G[v].size(); ++i) { edge& e = G[v][i]; if (e.capacity > 0 && dis[e.to] > dis[v] + e.cost + H[v] - H[e.to]) { dis[e.to] = dis[v] + e.cost + H[v] - H[e.to]; PreV[e.to] = v; PreE[e.to] = i; q.push(pair<int, int>(dis[e.to], e.to)); } } } if (dis[t] == INF)break; for (int i = 0; i <= V; ++i)H[i] += dis[i]; int d = f; for (int v = t; v != s; v = PreV[v])d = min(d, G[PreV[v]][PreE[v]].capacity); f -= d; flow += d; res += d*H[t]; for (int v = t; v != s; v = PreV[v]) { edge& e = G[PreV[v]][PreE[v]]; e.capacity -= d; G[v][e.rev].capacity += d; } } return res; } int Max_cost_max_flow(int s, int t, int f, int& flow) { int res = 0; fill(H, H + 1 + V, 0); while (f) { priority_queue <pair<int, int>> q; fill(dis, dis + 1 + V, -INF); dis[s] = 0; q.push(pair<int, int>(0, s)); while (!q.empty()) { pair<int, int> now = q.top(); q.pop(); int v = now.second; if (dis[v] > now.first)continue; for (int i = 0; i < G[v].size(); ++i) { edge& e = G[v][i]; if (e.capacity > 0 && dis[e.to] < dis[v] + e.cost + H[v] - H[e.to]) { dis[e.to] = dis[v] + e.cost + H[v] - H[e.to]; PreV[e.to] = v; PreE[e.to] = i; q.push(pair<int, int>(dis[e.to], e.to)); } } } if (dis[t] == -INF)break; for (int i = 0; i <= V; ++i)H[i] += dis[i]; int d = f; for (int v = t; v != s; v = PreV[v])d = min(d, G[PreV[v]][PreE[v]].capacity); f -= d; flow += d; res += d*H[t]; for (int v = t; v != s; v = PreV[v]) { edge& e = G[PreV[v]][PreE[v]]; e.capacity -= d; G[v][e.rev].capacity += d; } } return res; } }mcmf;
注:
参考链接:https://blog.csdn.net/u014800748/article/details/44059993
标签:cost 时间复杂度 blog 最大 顶点 art 函数 als bfs
原文地址:https://www.cnblogs.com/lfri/p/11267000.html