Clone Graph
Clone an undirected graph. Each node in the graph contains a label and a list of its neighbors.
OJ's undirected graph serialization:
Nodes are labeled uniquely.
We use # as a separator for each node, and , as a separator for node label and each neighbor of the node.
As an example, consider the serialized graph {0,1,2#1,2#2,2}.
The graph has a total of three nodes, and therefore contains three parts as separated by #.
First node is labeled as 0. Connect node 0 to both nodes 1 and 2.
Second node is labeled as 1. Connect node 1 to node 2.
Third node is labeled as 2. Connect node 2 to node 2 (itself), thus forming a self-cycle.
Visually, the graph looks like the following:
1
/ \
/ \
0 --- 2
/ \
\_/
Solution 1:
使用BFS来解决此问题。用一个Queue来记录遍历的节点,遍历原图,并且把复制过的节点与原节点放在MAP中防止重复访问。
图的遍历有两种方式,BFS和DFS
这里使用BFS来解本题,BFS需要使用queue来保存neighbors
但这里有个问题,在clone一个节点时我们需要clone它的neighbors,而邻居节点有的已经存在,有的未存在,如何进行区分?
这里我们使用Map来进行区分,Map的key值为原来的node,value为新clone的node,当发现一个node未在map中时说明这个node还未被clone,
将它clone后放入queue中处理neighbors。
使用Map的主要意义在于充当BFS中Visited数组,它也可以去环问题,例如A--B有条边,当处理完A的邻居node,然后处理B节点邻居node时发现A已经处理过了
处理就结束,不会出现死循环。
queue中放置的节点都是未处理neighbors的节点。
http://www.cnblogs.com/feiling/p/3351921.html
1 /* 2 Iteration Solution: 3 */ 4 public UndirectedGraphNode cloneGraph1(UndirectedGraphNode node) { 5 if (node == null) { 6 return null; 7 } 8 9 UndirectedGraphNode root = null; 10 11 // store the nodes which are cloned. 12 HashMapmap = 13 new HashMap (); 14 15 Queue q = new LinkedList (); 16 17 q.offer(node); 18 UndirectedGraphNode rootCopy = new UndirectedGraphNode(node.label); 19 20 // 别忘记这一行啊。orz.. 21 map.put(node, rootCopy); 22 23 // BFS the graph. 24 while (!q.isEmpty()) { 25 UndirectedGraphNode cur = q.poll(); 26 UndirectedGraphNode curCopy = map.get(cur); 27 28 // bfs all the childern node. 29 for (UndirectedGraphNode child: cur.neighbors) { 30 // the node has already been copied. Just connect it and don't need to copy. 31 if (map.containsKey(child)) { 32 curCopy.neighbors.add(map.get(child)); 33 continue; 34 } 35 36 // put all the children into the queue. 37 q.offer(child); 38 39 // create a new child and add it to the parent. 40 UndirectedGraphNode childCopy = new UndirectedGraphNode(child.label); 41 curCopy.neighbors.add(childCopy); 42 43 // Link the new node to the old map. 44 map.put(child, childCopy); 45 } 46 } 47 48 return rootCopy; 49 }
2014.12.30 Redo:
1 /* 2 SOLUTION 3: The improved Version. 3 */ 4 public UndirectedGraphNode cloneGraph(UndirectedGraphNode node) { 5 if (node == null) { 6 return null; 7 } 8 9 HashMapmap = new HashMap (); 10 11 // BUG 1: can't use queue , should use LinkedList. 12 Queue q = new LinkedList (); 13 14 q.offer(node); 15 16 // copy the root node. and then put it into the map. 17 UndirectedGraphNode nodeCopy = new UndirectedGraphNode(node.label); 18 map.put(node, nodeCopy); 19 20 while (!q.isEmpty()) { 21 UndirectedGraphNode cur = q.poll(); 22 23 // get out the copy node. We guarantee that it has been copied. Because we always put it into the map before 24 // put it into the queue. 25 UndirectedGraphNode curCopy = map.get(cur); 26 27 // go through all the children node. 28 // Line 71: java.util.ConcurrentModificationException. use cur instead of curCopy 29 for (UndirectedGraphNode child: cur.neighbors) { 30 31 if (map.containsKey(child)) { 32 curCopy.neighbors.add(map.get(child)); 33 } else { 34 // Only add the child into the map when it is not visited. 35 q.offer(child); 36 37 // BUG 3: forget to add the new node into the map. 38 UndirectedGraphNode childCopy = new UndirectedGraphNode(child.label); 39 curCopy.neighbors.add(childCopy); 40 map.put(child, childCopy); 41 } 42 } 43 } 44 45 return map.get(node); 46 }
Solution 2:
同样的,我们也可以使用递归DFS来解决此题,思路与上图一致,但为了避免重复运算产生死循环。当进入DFS时,如果发现map中已经有了拷贝过的值,直接退出即可。
题目虽然简单,但主页君仍然考虑了递归的特性使程序简洁。比如:我们拷贝只拷贝根节点,而子节点的拷贝由recursion来完成,这样可以使程序更加简洁。
注意:要先加入到map,再调用rec ,否则会造成不断地反复拷贝而死循环。
1 /* 2 Solution 2: Recursion version. 3 */ 4 public UndirectedGraphNode cloneGraph(UndirectedGraphNode node) { 5 if (node == null) { 6 return null; 7 } 8 9 return rec(node, new HashMap()); 10 } 11 12 public UndirectedGraphNode rec(UndirectedGraphNode root, HashMap map) { 13 // If it has been copied, just return the copy node from the map. 14 UndirectedGraphNode rootCopy = map.get(root); 15 if (rootCopy != null) { 16 return rootCopy; 17 } 18 19 // if the root is not copied, create a new one. 20 rootCopy = new UndirectedGraphNode(root.label); 21 map.put(root, rootCopy); 22 23 // copy all the child node. 24 for (UndirectedGraphNode child: root.neighbors) { 25 // call the recursion to create all the children and add the new children to the copy node. 26 rootCopy.neighbors.add(rec(child, map)); 27 } 28 29 return rootCopy; 30 }
2014.12.30 Redo:
1 public UndirectedGraphNode cloneGraph1(UndirectedGraphNode node) { 2 if (node == null) { 3 return null; 4 } 5 6 return rec(node, new HashMap()); 7 } 8 9 // SOLUTION 1: 10 // Try to return a copied cloneGraph. 11 public UndirectedGraphNode rec(UndirectedGraphNode node, HashMap map) { 12 // The base case: 13 if (map.containsKey(node)) { 14 // If the map has been copied, just return the node. 15 return map.get(node); 16 } 17 18 // create a new node. 19 UndirectedGraphNode nodeCopy = new UndirectedGraphNode(node.label); 20 // BUG 2: should put it into the map first. Because we don't want to copy the same node again in the recursion. 21 22 map.put(node, nodeCopy); 23 for (int i = 0; i < node.neighbors.size(); i++) { 24 // BUG 1: forget a parameter. 25 // copy all the children node. 26 nodeCopy.neighbors.add(rec(node.neighbors.get(i), map)); 27 } 28 29 30 31 return nodeCopy; 32 }
Ref: http://m.blog.csdn.net/blog/hellobinfeng/17497883
Code:
CloneGraph.java