Series Version: For a sequence {fk:N→[0,∞)}k∈N of functions with non-negative values
j=1∑∞k→∞liminffk(j)≤k→∞liminfj=1∑∞fk(j),∀j∈N
Proof
Strategy: It helps to have an intuition for simple functions. As with all lemmas, they are useful but the proof is quite long and complex, so it is best read when in a clear and healthy state of mind. The proof for the series case is essentially the same.
Part 1.
If we have f:=n→∞liminffngn:=k≥ninffk
then f=n→∞limgn.
∫Efdm≤n→∞liminf∫Efndm
To show this, it’s sufficient to prove for all simple functionsϕ≤f that
∫Eϕdm≤n→∞liminf∫Efndm
Now, let’s define a new simple function for a very small positive number ε>0. Then for a sufficiently large natural numbern, ϕε≤gn≤f will hold. Lastly, if we define Ak:={x∣gk≥ϕε},
Ak⊂Ak+1k=1⋃∞Ak=R
Part 2.
Since An and ϕε≤gn,
∫An∩Eϕεdm≤∫An∩Egndm
and since gn=k≥ninffk,
from k≥n, we have
∫An∩Egndm≤∫An∩Efkdm
Meanwhile, since An∩E⊂E,
∫An∩Efkdm≤∫Efkdm
Hence, we obtain the following:
∫An∩Eϕεdm≤n→∞liminf∫Efndm
Part 3.
Since ϕϵ was a simple function, its image can be represented as a finite set {c1,c2,⋯,cr}. If we assume Bi:=ϕϵ−1({ci}),
∫An∩Eϕεdm=i=1∑rcim(An∩E∩Bi)
From Part 1, since An⊂An+1 and n=1⋃∞An=R,
n→∞limi=1∑rcim(An∩E∩Bi)=n→∞lim∫An∩Eϕεdm=∫Eϕεdm
Combining with results obtained in Part 2,
∫Eϕεdm≤k→∞liminf∫Efkdm
Part 4.
Case 1. m({x∣ϕ(x)>0})<∞ ∫Eϕεdm=∫Eϕdm−εm({x∣ϕ(x)>0})
Taking the limit as ε→0, yields the following:
∫Eϕεdm=∫Eϕdm
Case 2. m({x∣ϕ(x)>0})=∞Dn:={xg(x)≥21min{ci}i=1n}
Then,
Dn⊂Dn+1n=1⋃∞Dn=R
Since ∫Dn∩Egndm→∞, and from the definition of gn, we obtain the following:
∫Dn∩Egndm≤∫Dn∩Efkdm≤∫Efkdm
Therefore, in any case, for all simple functionsϕ≤f, the following holds:
∫Eϕdm≤n→∞liminf∫Efndm
■
Capinski. (1999). Measure, Integral and Probability: p82. ↩︎