算法概述

递归算法简洁明了、可读性好，但与非递归算法相比要消耗更多的时间和存储空间。为提高效率，我们可采用一种非递归的二叉树遍历算法。非递归的实现要借助栈来实现，因为堆栈的先进后出的结构和递归很相似。

对于中序遍历来说，非递归的算法比递归算法的效率要高的多。其中序遍历算法的实现的过程如下：

(1).初始化栈，根结点进栈；

(2).若栈非空，则栈顶结点的左孩子结点相继进栈，直到null(到叶子结点时)退栈；访问栈顶结点(执行visit操作)并使栈顶结点的右孩子结点进栈成为栈顶结点。

(3).重复执行(2)，直至栈为空。

算法实现

package datastructure.tree;import datastructure.stack.ArrayStack;import datastructure.stack.Stack;public class UnrecOrderBTree implements Visit{	private Stack stack = new ArrayStack();	private BTree bt;	@Override	public void visit(BTree btree) {		System.out.print("\t" + btree.getRootData());	}		public void inOrder(BTree boot) {		stack.clear();		stack.push(boot);		while(!stack.isEmpty()) {			//左孩子结点进栈			while((bt = ((BTree)(stack.peek())).getLeftChild()) != null) {				stack.push(bt);			}			//如果该结点没有右孩子，则逐级往上出栈			while(!stack.isEmpty() &&!( (BTree)stack.peek() ).hasRightTree()) {				bt = (BTree)stack.pop();				visit(bt);			}			//如果该结点有右孩子，则右孩子进栈			if(!stack.isEmpty() && ( (BTree)stack.peek() ).hasRightTree()){				bt = (BTree)stack.pop();				visit(bt);				stack.push(bt.getRightChild());			}		}	}}

测试：

package datastructure.tree;/** * 测试二叉树 * @author Administrator * */public class BTreeTest {	public static void main(String args[]) {		BTree btree = new LinkBTree('A');		BTree bt1, bt2, bt3, bt4;		bt1 = new LinkBTree('B');		btree.addLeftTree(bt1);		bt2 = new LinkBTree('D');		bt1.addLeftTree(bt2);				bt3 =  new LinkBTree('C');		btree.addRightTree(bt3);		bt4 =  new LinkBTree('E');		bt3.addLeftTree(bt4);		bt4 =  new LinkBTree('F');		bt3.addRightTree(bt4);				RecursionOrderBTree order = new RecursionOrderBTree();		System.out.println("\n中序遍历：");		order.inOrder(btree);			}}