Joint Penn-UDel Colloquium on the Nature of Computing


NP-Complete via Railways & Molecular Computations

Maurice Margenstern

Groupe d'Informatique Fondamentale de Metz, the University of Metz and I.U.T. of Metz




Fig. 1. Going from A to B may turn out to be hard!



Universality and NP-Completeness in the Railways
(General Interest: First 25 minutes)

In 1994, Ian Stewart devised a railway circuit with only three kinds of switching points that is able to mimic the computation of a Turing machine. We give here a new proof of that universality result using a somewhat simpler circuit with the same kinds of switches. Both circuits, Stewart's and the author's are infinite. In finite circuits, we indicate a problem which is in NP-complete, using the same kind of switches. This explains why in today's life unexpected things sometimes happen with trains...

(Break: Second 10 minutes)

A Frontier Result on Molecular Computations
(Researcher's Interest: Last 25 minutes)

We shall first recall what molecular computations are. Then we explain what we mean by "frontier". We give an example of such a frontier result: time-varying distributed H-systems, TVDH, were devised to closer mimic natural processes that are performed according periodical cycles of operations. The degree of a TVDH is the number of operations in a cycle. The author and Yurii Rogozhin proved that starting from degree 2, it is always possible to devise TVDH's which generate any recursively enumerable language and consequently, have an unpredictable behavior, but for degree 1 this is not possible because TVDH of degree 1 generate regular languages, and so their behavior can be completely predicted.


At 3:30pm Thursday, September 16, at the University of Delaware, in Room 306 of Gore Hall at the east end of the pedestrian overpass across South College Street (near Amstel Avenue). Travel Directions are found at http://www.cis.udel.edu/~case/visitor_information.html.

The Joint Penn-UDel Colloquium on the Nature of Computing meets on each month's third Thursday. To receive future announcements, or make suggestions, send email to wood@cis.udel.edu or check our Joint Colloquium home page, http://www.cis.udel.edu/~wood/DNA/colloquium
Prior colloquium
Next colloquium
Home Page of Joint Colloquium

Maurice Margenstern: NP-Complete Railways & Molecular Computations
Compiled by / wood@cis.udel.edu / Last revised September 24, 1999