Consider array A of n numbers. We want to design a dynamic programming algorithm to find the maximum sum of consecutive numbers with at least one number. Clearly, if all numbers are positive, the maximum will be the sum of all the numbers. On the other hand, if all of them are negative, the maximum will be the largest negative number. The complexity of your dynamic programming algorithm must be O(n²). However, the running time of the most efficient algorithm is O(n). Design the most efficient algorithm you can and answer the following questions based on it. To get the full points you should design the O(n) algorithm. However, if you cannot do that, still answer the following questions based on your algorithm and you will get partial credit. (a) tion(s) to subproblem(s). Briefly explain how it computes the solution. Do not forget the base case(s) of your recursion. Write the recursion that computes the optimal solution recursively based on the solu-

Suppose that each row of an n × n array A consists of 1's and 0's such that in any row of A all the 1's come before any O's in that row. Assuming that A is already in memory, describe an algorithm running in O(n) time for finding the row that contains the most 1's. Write the algorithm following the style of your course-notes.

Suppose that you have five different algorithms (A, B, C, D and E) for solving a problem. To solve a problem of size n, the number of operations used by each algorithm is as follows. As n grows, which algorithm uses the most operations? Algorithm number of operations n2+n3 B 2n+1 log n' 26+ log n n(logn)? +1 n100 Algorithm D Algorithm C Algorithm A Algorithm B o Algorithm E

1. Asymptotics. Given an array A of n integers, you'd like to output a two-dimensional n x n array B in which B[i, j] = max {A[i], A[i + 1],..., A[j]} for each i < j. For i j the value of B[i, j] can be left as is. for i = 1,2, η for j = i + 1, n " Compute the maximum of the entries A[i], A[i + 1], , A[j]. Store the maximum value in B[i, j]. (a) Find a function of such that the running time of the algorithm is O(f(n)), and clearly explain why. (b) For the same function f argue that the running time of the algorithm is also (f(n)). (This establishes an asymptotically tight bound (f(n)).) (c) Design and analyze a faster algorithm for this problem. You should give an algorithm with running O(g(n)), where lim→∞ g(n)/f(n) = 0.