Divide and Conquer Parallel and Sequential Data Structures

Let’s look at an example. For s = 〈1,−5, 2,−1, 3〉, we know that 〈1〉, 〈2,−1, 3〉, and 〈−5, 2〉 are all contiguous subsequences of s—whereas 〈1, 2, 3〉 is not. Among such subsequences, we’re interested in finding one that maximizes the sum. In this particular example, we can check that mcss(s) = 4, achieved by taking the subsequence 〈2,−1, 3〉. We have to be careful about what our MCSS problem returns given an empty sequence as input. Here we assume the max of an empty set is −∞ so it returns −∞. We could alternatively decide it is undefined.