The first few are to be submitted to the drop box and the last one committed to your repository. If either part (dropbox or repo) is late, you will be charged a late day. If both parts are early, you can gain a late day
(20 points) Calculate the exact average-case time complexity of binary search. (You can use any standard implementation of binary search, it's in Figure 2.2.1 of our zybook text and in Wikipedia, for example) Only consider successful searches (where the item we are looking for is actually in the array) and assume all elements are equally likely to be the one we are searching for.
Hint: For each i, count how many of the array elements will be found after binary search examines exactly i array elements. (For example, the middle element is the only one to be found after examining one array element, and there are two elements that can be found after examining two array elements.) Then sum over all and divide to find the average. For the generalization, you'll use a summation. You need to convince me somehow that your summation is correct, perhaps using a table or words. Once you have done that, then you may find the following formula useful:
After you have done that, you'll get a messy expression with lots of terms in it. Take the limit as n gets large so some terms go to zero. You'll end up with a much simpler expression that should be clearly between 1 and log(n), like we said in class.
(Hint: that section of the textbook gives an answer for the worst-case runtime of binary search, so you know that your answer should be smaller than that. But how much smaller is the interesting question we are trying to answer here.