home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.math
- Path: sparky!uunet!ccorp!keithw
- From: keithw@ccorp.uucp (Keith Weintraub - dpr2)
- Subject: Calculus of Variations?
- Message-ID: <1992Jul23.212420.8636@ccorp.uucp>
- Organization: Citicorp, New York City
- Date: Thu, 23 Jul 92 21:24:20 GMT
- Lines: 32
-
- Netfolk,
- I am working on forward rate curve construction and would like to
- solve a problem that I think is of "Calculus of Variations" type. Feel free
- to correct me if I am wrong. I would like to find a *continuous* non-negative
- function y(t) whose derivative has minimum L_2 norm over [0,T] and integrates
- to given constants over disjoint subintervals that cover [0,T]. (I wanted to
- write that in English---hope I succeeded---in case people don't like the TeX
- notation below). More symbolically I would like to find a *continuous*
- function y(t) such that
-
- \int_0^T (y'(s))^2 ds
-
- is a minimum subject to the constraints
-
- \int_{t_{i-1}}^{t_i} y(s) ds = a_i > 0 (a_i known) i = 1,...,n
-
- where
- 0 = t_0 < t_1 < ... < t_n = T.
-
- Any guidance, references, solutions would be most appreciated.
-
- Respond by e-mail or post and if I get anything interesting I will post a
- summary.
-
- Thanks in advance,
- KW
-
-
- --
- Keith Weintraub (KW) -- Citicorp | I told you a trillion times:
- keithw@Citicorp.COM | "Don't exaggerate."
- uunet!ccorp!keithw |
-