At the University of Kaiserslautern , I worked with Juergen Paulokat and Barbara Dellen on the same issues as previously reported. I now have a better understanding of Paulokat's current work, which I hope to relay to Heecheol for the Next-Link work. Juergen notes that though his constraint propagation system is written in Smalltalk, a C "wrapper" is available and that he hopes to be able to use the KAPI code to make his code available as a Next-Link agent.
My host was Frank Maurer. The Software Engineering project, reported as proposed in the previous report, has been funded. I also met with Sigrid Goldman, a masters student, who may do her work at Stanford on the Next-Link project. She has been working in Software Engineering and is interested in process modeling. I attended and gave the technical keynote at the XPS-95 conference. Despite slide printing problems, it went very well. These slides will be used to upgrade the Next-Link web pages and for future talks at Stanford.
I revisited Ian Smith at the AI Lab in the Swiss Federal Institute of Technology (EPFL) and worked with him and Esther Gelle on general backtracking issues in planning. Esther and I did a brief analysis of the bridge design problem. We agreed that a more explicit control structure is needed, perhaps along the lines of the general Redux model. We ftp'd the Redux' code and recompiled it under Allegro lisp. This would be one component of the eventual general system, also including a top level control of goals, a constraint propagation system to suggest and evaluate parameter values, a sensitivity analysis component to suggest the best way to backtrack, and Redux' to control search.
I also saw a demonstration of Berthe Yazid Choueiry's work on task/resource assignments to a pre-existing schedule. The demonstration I saw was on hospital personel assignments, but the technique would be equally applicable to the possible application to Swiss Air as well as many other possibilites. The technique of "bundling" variable domains was very impressive.
I also met with Henk Slettenhaar to work on the Swiss study tour for June. Ian, Henk, and I were given a tour of the EPFL manufacturing cell by Prof. Jean-Dominique Decontignie@di.efl.ch who is looking for a sponsor for a visit to Stanford. I have his papers.
The actual travel plan is below.
Gave a keynote at the
XPS-95 on Friday 3 Mar 95.
My apartment number is
49 631 31615 36
Also:
Contact:
Juergen Paulokat
University of Kaiserslautern
P. O. Box 3049
D-67653 Kaiserslautern, GERMANY
hm 49 631 17416
off 49 631 2053367
paulokat@uklirb.informatik.uni-kl.de