Determine the greatest possible value of for real numbers satisfying .
Solution
The maximum value is . Since , it is equivalent to maximize for with ; note that this domain is compact, so the maximum value is guaranteed to exist. For convenience, we establish something slightly stronger: we maximize for with , where may be any even nonnegative integer up to , and show that the maximum is achieved when . We first study the effect of varying and while fixing their sum. If that sum is , then the function has constant second derivative , so it is either everywhere convex or everywhere concave. Consequently, if achieves the maximum, then for any two indices , at least one of the following must be true: one of , is extremal (i.e., equal to or ); (in which case and the local maximum is achieved above); (in which case above). In the third case, we may discard and and achieve a case with smaller ; we may thus assume that this does not occur. In this case, all of the non-extremal values are equal to some common value , and moreover we cannot have both 1 and -1. We cannot omit 1, as otherwise the condition cannot be achieved; we must thus have only the terms 1 and , occurring with some positive multiplicities and adding up to . Since and , we can solve for to obtain ; we then have Since , we must have . For fixed , the target function increases as increases, so the optimal case must occur when . The possible pairs at this point are computing the target function for these values yields respectively yielding as the maximum value.