Interview-Increasing Sequence with Length 3.
Given an array, determine whether there are three elements A[i],A[j],A[k], such that A[i]<A[j]<A[k] & i<j<k.
Analysis:
It is a special case of the Longest Increasing Sequence problem. We can use the O(nlog(n)) algorithm, since we only need sequence with length three, we can do it in O(n).
Solution:
1 public static boolean threeIncSeq(int[] A){ 2 if (A.length<3) return false; 3 4 int oneLen = 0; 5 int twoLen = -1; 6 for (int i=1;i<A.length;i++){ 7 //check whether current element is larger then A[twoLen]. 8 if (twoLen!=-1 && A[i]>A[twoLen]) return true; 9 if (twoLen!=-1 && A[i]>A[twoLen] && A[i]<A[oneLen]){ 10 twoLen = i; 11 continue; 12 } 13 if (twoLen==-1 && A[i]>A[oneLen]){ 14 twoLen = i; 15 continue; 16 } 17 if (A[i]<A[oneLen]){ 18 oneLen = i; 19 continue; 20 } 21 } 22 23 return false; 24 }
Variant:
If we need to output the sequence, we then need to record the sequence of current length 1 seq and length 2 seq.
1 public static List<Integer> threeIncSeq(int[] A){ 2 if (A.length<3) return (new ArrayList<Integer>()); 3 4 int oneLen = 0; 5 int twoLen = -1; 6 List<Integer> oneList = new ArrayList<Integer>(); 7 List<Integer> twoList = new ArrayList<Integer>(); 8 oneList.add(A[0]); 9 for (int i=1;i<A.length;i++){ 10 //check whether current element is larger then A[twoLen]. 11 if (twoLen!=-1 && A[i]>A[twoLen]){ 12 twoList.add(A[i]); 13 return twoList; 14 } 15 if (twoLen!=-1 && A[i]>A[twoLen] && A[i]<A[oneLen]){ 16 twoLen = i; 17 twoList = new ArrayList<Integer>(); 18 twoList.addAll(oneList); 19 twoList.add(A[i]); 20 continue; 21 } 22 if (twoLen==-1 && A[i]>A[oneLen]){ 23 twoLen = i; 24 twoList = new ArrayList<Integer>(); 25 twoList.addAll(oneList); 26 twoList.add(A[i]); 27 continue; 28 } 29 if (A[i]<A[oneLen]){ 30 oneLen = i; 31 oneList = new ArrayList<Integer>(); 32 oneList.add(A[i]); 33 continue; 34 } 35 } 36 37 return (new ArrayList<Integer>()); 38 39 }
NOTE: This is more compliated then needed, when using List<> in this case, but this idea can be used to print the LIS.