Wednesday, 27 January 2010

measure theory - Generalisation of Lebesgue decomposition theorem

A little further searching turned up a simple proof of Lebesgue's decomposition theorem in "The Lebesgue Decomposition Theorem for Measures", J. K. Brooks, The American Mathematical Monthly, 78 (1971), pp. 660-662. Without much extra work, it admits the following generalisation.



Theorem. Let mathcalNsubsetOmega be a collection of subsets such that



  1. if EinmathcalN and FinOmega, FsubsetE, then FinmathcalN;

  2. if EninmathcalN is a countable collection, then bigcupnEninmathcalN as well.

Consider the subspace mathcalS=lbracemumidmu(E)=0textforallEinmathcalNrbrace. Then mathcalM=mathcalSoplusmathcalSperp.



Proof. Fix nuinmathcalM, and consider the following collection of subsets:
mathcalR=lbraceEinmathcalNmidnu(E)>0rbrace.


Let alpha=suplbracenu(E)midEinmathcalRrbrace, and let EninmathcalR be a sequence of sets such that nu(En)toalpha. Let A=bigcupnEn. Then nu(A)=alpha and AinmathcalN.



Furthermore, given any EinmathcalR, we have nu(EsetminusA)=0. Indeed, if nu(EsetminusA)>0, then nu(E)=nu(A)+nu(EsetminusA)>alpha, contradicting the definition of alpha. Similarly, nu(EsetminusA)=0 for every EinmathcalN.



Thus we may take nu1=nu|XsetminusA and nu2=nu|A. It follows that nu2inmathcalSperp, since nu2(A)=1 and AinmathcalN, and nu1inmathcalS, since nu(EsetminusA)=0 for every EinmathcalN.



Finally, uniqueness follows since mathcalScapmathcalSperp=lbrace0rbrace.

No comments:

Post a Comment