Integration on a nice enough manifold of a function $f:M \to \mathbb{R}$ is defined $$\int f = \sum_{ i \in I} \int_{U_i}\phi_i f$$ where $\phi_i$ is a partition of unity subordinate to the open cover $U_i$ of $M$.

The index set $I$ can be infinite in general.

If $M$ is compact (bounded), can I take $I$ to be a finite set so that the sum is a finite sum?

  • 3
    $\begingroup$ Yes, if $M$ is compact, you can always have a finite partition of unity subordinate to any open cover. $\endgroup$ – Daniel Fischer Jan 12 '14 at 23:37

Yes, since every open cover has a finite subcover (Heine-Borel).

  • $\begingroup$ Thanks, but why don't books emphasise that the sum can be finite?? $\endgroup$ – michael_carbon Jan 12 '14 at 23:39
  • $\begingroup$ I am sure some do. Also, many manifolds are not compact, but they are generally locally compact, which means that every point lies in a finite number of the open sets (this is obviously important, since infinite sums are a big pain). $\endgroup$ – Igor Rivin Jan 12 '14 at 23:40
  • $\begingroup$ @PedroTamaroff I actually think that the local finiteness (global finiteness less so) is crucial for the whole machine to work, and I am sure if you read the first partitions of unity papers, it is probably emphacized. $\endgroup$ – Igor Rivin Jan 12 '14 at 23:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.