I have the following implementation of Kadane's algorithm in java. It is basically to find the maximum sum of contiguous subarray.
String[] numbers = string.split(",");
int max_so_far = 0;
int max_ending_here = 0;
for (int i = 0; i < numbers.length-1;i++){
max_ending_here = max_ending_here + Integer.parseInt(numbers[i]);
if (max_ending_here < 0)
max_ending_here = 0;
if (max_so_far < max_ending_here)
max_so_far = max_ending_here;
}
System.out.println(max_so_far);
However this doesn't work if there is a combination of a negative and positive number in an array, for example the following:
2,3,-2,-1,10
Which should return a 12 as a maximum. As of now it returns 5
You algorithm implementation looks ok, but your loop conditional i < numbers.length-1
does not: it stops just 1 short of the end of the array. i < numbers.length
should do it :-)
this works for me:
String string = "2,3,-2,-1,10";
String[] numbers = string.split(",");
int max_so_far = 0;
int max_ending_here = 0;
for (String num : numbers) {
int x = Integer.parseInt(num);
max_ending_here = Math.max(0, max_ending_here + x);
max_so_far = Math.max(max_so_far, max_ending_here);
}
System.out.println(max_so_far);
Regarding the above answer by Michał Šrajer:
Line #7: max_ending_here = Math.max(0, max_ending_here + x);
should be:
max_ending_here = Math.max(x, max_ending_here + x);
...according to the Kadane algorithm as defined here
Too late, but if someone needs it in the future.
public static void kadaneAlgo(int[][] array)
for(int i = 1; i < array.length; i++){
int max_value_index_i = numberOrSum(array[i], past);
if(max_value_index_i > sum){
sum = max_value_index_i;
}
past = max_value_index_i;
}
System.out.println("Max sum from a contiguous sub array is : " + sum);
}
public static int numberOrSum(int number, int past){
return Math.max(number, number+past);
}