求字符串中不含重复字符的最长子串的长度

原帖：http://topic.csdn.net/u/20100930/02/5064ef5a-b2f4-44b9-8a16-ce2bf83323c0.html?84726

题目：求字符串最长不含重复字符的子串长度，如abcbec，就返回3.

分析：

利用动态规划(DP)原理，设字符串S的长度为n，考虑i...n-1这个后缀中符合条件的子串：首先需要记录两组数据，第一组数据是从i向右找到的最长不含重复字符的子串长度prefixlen[i]，第二组数据是在i...n-1后缀中符合条件的子串之起始和结束位置，分别用maxlenstart[i]和maxlenend[i]表示，注意二组数据满足：maxlenstart[i]-maxlenend[i]+1>=prefixlen[i]，即从S[i]开始最左边的最长不含重复字符的子串长度与区间i...n-1中符合条件的子串长度可能相等；如果S[i]给prefixlen[i+1]增加一个字符长度，说明S[i]与右边长度为prefixlen[i+1]个字符都不相同，这时需要分两种情况考虑，1）如果maxlenstart[i+1]==i+1即S[i+1...n-1]中得到的最优子串的左边一个字符与S[i]也相邻，说明最大子串长度也应该增加1, 即设置maxlenstart[i]=i；注意此时如果prefixlen[i]总长度未在prefixlen[i+1]的基础上加1,说明第二组数据的长度不会比第一组数据长度大了，这是因为根据DP的计算过程，只要第二组数据与S[i]相邻，那么两组数据长度保证了前一步中是相等的；2）如果第二组数据不与S[i]相邻，则需要比较二组数据的长度了。

具体实现时需要注意边界条件，设置prefixlen[n] = 0; maxlenstart[n] = n; maxlenend[n] = n-1;。

代码：

/**
 * 
 * @author ljs
 * 2011-07-09
 *
 */
public class LongestSubseqWithUniqueChars {

	public static int solve(String str){
		int n = str.length();
		int[] prefixlen=new int[n+1];
		prefixlen[n] = 0;
		int[] maxlenstart = new int[n+1];
		maxlenstart[n] = n;
		int[] maxlenend = new int[n+1];
		maxlenend[n] = n-1;
		
		for(int i=n-1;i>=0;i--){
			char c = str.charAt(i);
			//caculate prefixlen[i] by prefixlen[i+1]
			int j=0,k;
			for(j=0,k=i+1;j<prefixlen[i+1];j++,k++){
				if(c == str.charAt(k)){					
					break;
				}
			}
			prefixlen[i] = j+1;
			
			//caculate maxlenstart[i] and maxlenend[i]
			maxlenstart[i] = maxlenstart[i+1];
			maxlenend[i] = maxlenend[i+1];
			if(maxlenstart[i+1] == i+1){
				if(prefixlen[i] == prefixlen[i+1] + 1){
					maxlenstart[i] = i;
				}
			}else{
				//update the max len for i...n-1
				if(maxlenend[i] - maxlenstart[i] + 1 < prefixlen[i]){
					maxlenstart[i] = i;
					maxlenend[i] = i + prefixlen[i] - 1;
				}
			}			
		}
		return maxlenend[0] - maxlenstart[0] + 1;
	}
	public static void main(String[] args) {
		String str = "abcbec";
		int maxlen = LongestSubseqWithUniqueChars.solve(str);
		System.out.format("max len of substring with unique chars for string \"%s\" = %d.%n",
				str,maxlen);
		
		str = "adabcbec";
		maxlen = LongestSubseqWithUniqueChars.solve(str);
		System.out.format("max len of substring with unique chars for string \"%s\" = %d.%n",
				str,maxlen);
		
		str = "abadadabbc";
		maxlen = LongestSubseqWithUniqueChars.solve(str);
		System.out.format("max len of substring with unique chars for string \"%s\" = %d.%n",
				str,maxlen);
		
		str = "ffdeefghff";
		maxlen = LongestSubseqWithUniqueChars.solve(str);
		System.out.format("max len of substring with unique chars for string \"%s\" = %d.%n",
				str,maxlen);
		
		str = "abcbecghijkl";
		maxlen = LongestSubseqWithUniqueChars.solve(str);
		System.out.format("max len of substring with unique chars for string \"%s\" = %d.%n",
				str,maxlen);
		
	}

}

测试输出：

max len of substring with unique chars for string "abcbec" = 3.
max len of substring with unique chars for string "adabcbec" = 4.
max len of substring with unique chars for string "abadadabbc" = 3.
max len of substring with unique chars for string "ffdeefghff" = 4.
max len of substring with unique chars for string "abcbecghijkl" = 9.