Features
-
Only public domain software packages that were tested and found to be
workable and reasonably efficient and easy to operate will be made available
to users of RIOT. Some information about the software characteristics and
performance (e.g. size limitation, speed) will be made available.
-
Users will be able to search for the appropriate software by
searching through tree-structured taxonomy of optimization problems (e.g. max
flow, linear programming). However, an important feature of the proposed site
is that users who are not familiar with the jargon of optimization research
will be able to interactively identify the proper software through a dialog
provided by electronic forms.
-
To avoid time consuming transfer and installation of software packages,
users will be able to test samples of their problems remotely on the proposed
www site. Several choices regarding problem data structure will be available
including some popular modeling languages and spread sheets. When comparing
several software packages the users need only submit one set of problem data
which will be automatically transformed to the data structure required by the
different packages.
-
Optimization algorithms and applications developed within the
University of California
at Berkeley will be maintained and available for distribution on the
proposed site. Software developed elsewhere will be available through
links to appropriate www distribution sites.
-
The results of samples submitted by users (and their comments)
will be used to update the data base
recording the performance of the software packages. We anticipate that once
a critical mass is reached, the proposed www site will become the primary
place for optimization software developers to showcase their products as well
as the primary source of optimization algorithms for end users.
- For users unfamiliar with the algorithms covered, educational pages will
be available that describe basic concepts.
- Interative applications will be developed which will showcase the utility
of select algorithms.
Access statistics for this site are
available on-line.