This is an old revision of the document!
Lecture 3
Today we continue going over Tao's sequence of Lemma 7.4.2 - 7.4.11
Lemma 7.4.2
I will prove the easy case (1-dim), you will do the general case in HW.
Let E=(0,+∞) be the open half space in R. For any subset A⊂R, we need to prove that
m∗(A)=m∗(A∩Ec)+m∗(A∩E).
Let A+=A∩E and A−=A∩Ec, note that if 0∈A, then 0∈A−. First, we show that m∗(A)≤m∗(A−)+m∗(A+). This is because any open cover of A− and open cover of A+, union together form an open cover of A. Next, we need to show that, for any ϵ>0, we have
m∗(A)+2ϵ≥m∗(A−)+m∗(A+).
The plan is the following. Given an open covering of A by intervals {Bj}j=1∞, such that m∗(A)+ϵ>∑j∣Bj∣. We define
Bj+=Bj∩(0,∞),Bj−=Bj∩(−∞,ϵ/2j)
then {Bj+} forms a cover of open interval of A+, similarly {Bj+} forms a cover of open interval of A−. ∣Bj+∣+∣Bj−∣≤∣Bj∣+ϵ/2j.
Thus, we have
m∗(A)+2ϵ≥j∑∣Bj∣+ϵ≥j∑∣Bj+∣+j∑∣Bj−∣≥m∗(A+)+m∗(A−).
That finishes the proof for n=1 case. How to generalize to higher dimension? In the above discussion, we used the trick of 1=∑j1/2j, which is same trick as we prove outer-measure has countable sub-additivity. We can prove the general n case in two steps, first treat the case of A being an open box, then for general subset A⊂Rn, and given an open box cover {Bj} of A with ∑j∣Bj∣<m∗(A)+ϵ, we have
$$ m^*(A) + \epsilon > \sum_j |B_j| = \sum_j m^*(B_j) = \sum_j m^*(B_j \cap E) + m^*(B_j \RM E) \geq m^*(\cup_j(B_j \cap E)) + m^*(\cup(B_j \RM E)) \gep m^*(A \cap E) + m^*(A \RM E) $$