next up previous
Next: Signed Additive Functionals Up: Corollaries of Kac's Moment Previous: Corollaries of Kac's Moment

Positive Continuous Additive Functionals

  To this point our study has focused on the random variable tex2html_wrap_inline2808 , which is the value at time t=T of the additive functional tex2html_wrap_inline3290 . In the abstract, a positive continuous additive functional (PCAF) is an tex2html_wrap_inline3190 -adapted family tex2html_wrap_inline3294 of positive finite random variables satisfying the additivity condition

  equation1852

We can define an operator tex2html_wrap_inline3298 analogous to the operator tex2html_wrap_inline3300 by the formula

  equation1858

for positive tex2html_wrap_inline2864 -measurable f. Note that tex2html_wrap_inline3306 when tex2html_wrap_inline3308 . The validity of the analog of (13), namely

  equation1865

requires a mild additional hypothesis. For instance, if E is a complete separable metric space (with Borel tex2html_wrap_inline2862 -algebra tex2html_wrap_inline2864 ) and X has right-continuous sample paths, then (21) is valid with f and tex2html_wrap_inline3180 as for (13). The proof of this assertion involves Ray-Knight compactification methods found in [59], and is well beyond the scope of this article. Assuming the validity of (21), we can repeat the earlier argument to show that Kac's moment formula (4) and the F-K formula (7) hold for any PCAF A, provided the operator tex2html_wrap_inline3298 is substituted for tex2html_wrap_inline3300 . Ray [55], Section 2, used this version of the F-K formula for finite linear combinations of local times.



Patrick Fitzsimmons
Wed May 17 08:50:36 PDT 2000