d) Let P = First compute a spectral factorization P= QAQT as in Q.li). This will show that P is positive definite. Now suggest a method to draw & (P). For this, observe x Px = (@Tx)"A(Qx). How would this help you ? [Hint: When you use x = 91 or x=92, the first or second column of Q in your computations, what do you get ?] e) Now given a general n x n positive definite matrix P, based on what you have found out in part d), describe the shape of & (P). f) Suppose P is an n x n positive definite matrix and c ER". Consider the convex optimization problem minimize cx subject to xPx