My question is number 3.

3. Same as Problem 2, for a limited amount of foreknowledge. Namely, suppose we could know in advance whether to s 1 or ti > 1, for each i. Our schedule would then be in what is called ESPT (expected shortest processing time) order. We'd run all the jobs for which t; S 1, and then all the jobs for which ti > 1. Calculate the expected flowtime and compare it with both your answer to 2 and (190). What do you conclude about the value of partial information compared with the value of detailed information? Hint: if this problem is too hard, first solve the simpler problem where exactly 50 of the jobs have runtime 1. 2. Suppose job i hast; ~U (0,2) for i = 1...99 and the t, are jointly independent. If the jobs are run in order 1, 2, 3,...,99 on one processor, the expected flowtime is obviously , i = (99)(100)/2. On the other hand, if we could foresee the future, and know in advance the values attained by t., too, we could run the jobs in SPT order. How helpful would it be to foresee the future, with respect to flowtime? That is, what is the expected flowtime if the jobs are run in SPT order, and how does it compare with (120)? 3. Same as Problem 2, for a limited amount of foreknowledge. Namely, suppose we could know in advance whether to s 1 or ti > 1, for each i. Our schedule would then be in what is called ESPT (expected shortest processing time) order. We'd run all the jobs for which t; S 1, and then all the jobs for which ti > 1. Calculate the expected flowtime and compare it with both your answer to 2 and (190). What do you conclude about the value of partial information compared with the value of detailed information? Hint: if this problem is too hard, first solve the simpler problem where exactly 50 of the jobs have runtime 1. 2. Suppose job i hast; ~U (0,2) for i = 1...99 and the t, are jointly independent. If the jobs are run in order 1, 2, 3,...,99 on one processor, the expected flowtime is obviously , i = (99)(100)/2. On the other hand, if we could foresee the future, and know in advance the values attained by t., too, we could run the jobs in SPT order. How helpful would it be to foresee the future, with respect to flowtime? That is, what is the expected flowtime if the jobs are run in SPT order, and how does it compare with (120)