Solve the problem Please!!!!!

Problem 8. (40 POINTS) Google ranks web-pages with an arbitrary numerical value. One cannot make any assumptions about the range of the rank. Suppose that a query returns n pages ( n is part of the input) and information about rank is stored for the i-th page in A[i]. We are interested in determining the top 20%, the bottom 40% and the middle 40% of all pares. For that we need to determine two values m and M, the values for which 40% of the pages have ranks less than m,20% more than M and those whose rank is between m and M. Give an efficient algorithm (whose performance cannot be improved upon) that solves this problem, i.e. it finds m, and M given n and A