Take the last process -> it will get its first request processed after 10 schedules. Because after the first time quantum, the first process does an I/O operation (as it would have finished its request which also takes 20ms and hence CPU given to next process) and hence the second process gets its chance. Thus, after 9 time quantum, the final process gets its request served (assuming a previously given process is never given CPU time again until all other processes in wait are given). So, this means we need 10 schedules (including one at the beginning) and given in question that each scheduling takes 2ms. So, response time for the first request of the last process
$= 9\times 20 ms + 10 \times 2$
$= 180 + 20 = 200 ms$.
(2) The scheduling goes like this:
P1 P2 P3 ... P10 P1 P2 P3 ... P10 .... P10 P1 P2 ... P10 (P1-P10 repeated 15 times)
After every time quantum and at the beginning we need 2ms for scheduling the next process. So, total scheduling length
$= (20ms + 2ms)\times 10 \times 15$
$=220 \times 15 = 3300 ms$