home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!charon.amdahl.com!pacbell.com!sgiblab!darwin.sura.net!zaphod.mps.ohio-state.edu!rpi!crdgw1!rdsunx.crd.ge.com!barbados!ramesh
- From: ramesh@barbados.crd.ge.com (Rajaram Ramesh)
- Newsgroups: sci.math
- Subject: Help needed (Problem in integer programming?)
- Message-ID: <1992Nov18.184914.13989@crd.ge.com>
- Date: 18 Nov 92 18:49:14 GMT
- Sender: usenet@crd.ge.com (Required for NNTP)
- Organization: GE Corp. Research & Development, Schenectady, NY
- Lines: 40
- Nntp-Posting-Host: barbados.crd.ge.com
-
-
- I have come across a problem in my research which probably involves
- integer programming. I am looking for any insight into how to go about
- tackling this problem.
-
- Given a sequence f(n), n=0,...,M-1.
- Given:
- 1. f(0)=0;
- 2. f(k)=f(M-k), k >0, k<M
- 3. f(n) is a positive integer for n > 0.
- 4. an integer L < M
- If it is of any help, f(M/2)=M/2 and f(k) < M/2 for all other k.
-
- I wish to find a subset of indices k_1, k_2, ... k_L,
- that
- minimizes the double summation
-
- sum_{l=1}^L sum_{m=1}^L f((k_l - k_m) mod M)
-
- ,i.e.,
-
- f((k_1 - k_1) mod M) + f((k_1 - k_2)mod M) +... + f((k_1 - k_L) mod M) +
- f((k_2 - k_1) mod M) + f((k_2 - k_2)mod M) +... + f((k_2 - k_L) mod M) +
- .
- .
- f((k_L - k_1) mod M) + f((k_L - k_2)mod M) +... + f((k_L - k_L) mod M)
-
-
-
- I do not have to necessarily find the optimum solution,
- good sub-optimal solutions are okay.
-
-
- On the surface, it looks as if a complete search has to be done. If
- anyone knows of a trick to avoid this, I'd appreciate it.
-
- Thanks.
-
- Ramesh.
- ramesh@crd.ge.com
-