You should also check out a short note by Barzilai and co-authors
(writer rummages around in in stacks of papers but comes up empty-handed)
in OR Letters (ca. 1990). They modify the AHP by using geometric means
instead of Saaty's arbitrary eigenvectors and show that their method is
not subject to rank reversal. It should also be damned near trivial to
program, although I have not come across any implementation.

Hope that helps.

gus gassmann
gassm...@earth.sba.dal.ca

Hallo everybody,

I'm an analyst working on methods for R&D - project selection. One
possible approach should be Saaty's analytic hierarchy process (AHP).

i'm looking for

1.  newer literature about R&D - project selection using AHP.
2.  any (share) software for application of AHP.

Thanks in advance,

Markus Schmitt, Univ. of Berne, Switzerland

Two software packages suitable for R&D project selection are available
(for a fee) from Krysalis Ltd of England. The packages are named Hiview
and Equity. Hiview is a multiattribute utility model that facilitates use
of a hierarchy of attributes. It was recently reviewed in OR/MS Today.
Equity is a resource allocation model that helps one establish the
appropriate sizes for several projects that will all be funded
simultaneously. Unlike linear programming, Equity models each project as
a series of discrete alternatives for funding; you analyze the projects
to discover the right balance of limited funds across the projects.
Your point of contact for Krysalis Ltd is Colin Simmons, whom you can
reach at 100333.3...@compuserve.com
\Ken Kuskey

: Hallo everybody,

: I'm an analyst working on methods for R&D - project selection. One
: possible approach should be Saaty's analytic hierarchy process (AHP).

: i'm looking for

: 1.  newer literature about R&D - project selection using AHP.
: 2.  any (share) software for application of AHP.

: Thanks in advance,

: Markus Schmitt, Univ. of Berne, Switzerland

--
Ken Kuskey
2133 Golf Course Drive
Reston, VA 22091
703 476-5741 (voice)