The Generalized Minkowski Inequality
[5/40%]Generalized Minkowski Inequality: Let and denote -finite measure spaces, let , and let be -measurable and such that
Then
Consider the -finite measure space and consider the finite measure space , where , , and denotes counting measure. Let , where . Define by writing and for each . Here, the Generalized Minkowski Inequality becomes as follows: Let . Then and .
Let . Without worrying about measurability conditions, obtain the inequality
We have . Play with exponents and use the Generalized Minkowski Inequality.
Let denote the Lebesgue subsets of , and let denote Lebesgue measure on the Lebesgue subsets of . Consider the -finite measure space . Given a function , associate the function given by .
For we have and .
The constant is the best [smallest] possible constant for the inequality in (a).
We have excluded the case so far. Show that if is positive, then .
(a) Replace by , and then use the main result of this section.
(b) Consider for each .
This exercise is a lemma for the next exercise. Let denote a measure space with . Every -measurable function satisfies
Suppose is not essentially bounded. Then has positive -measure for each . Show for each and that
by splitting the integral into the integral over the set of where and the set of where . Next, since . Therefore, given any large , we may choose so large that exceeds, say, . This gives the result when is not essentially bounded. If on the other hand is essentially bounded, then for every by the fact that . We have shown earlier that in this case; see Exercise 4 in Section 8.5.
We proved the Generalized Minkowski Inequality for . There is a version for the case . Let and denote -finite measure spaces, and let be -measurable.
The mapping is -measurable, hence we may meaningfully discuss .
From now on, assume that .
For -a.e. the mapping is in . That is, for -a.e. we have .
The [nonnegative] mapping defined on is -measurable, hence we may meaningfully discuss .
We have the inequality