M. Tambe et Ps. Rosenbloom, INVESTIGATING PRODUCTION SYSTEM REPRESENTATIONS FOR NON-COMBINATORIALMATCH, Artificial intelligence, 68(1), 1994, pp. 155-199
Citations number
82
Categorie Soggetti
Computer Sciences, Special Topics","Computer Science Artificial Intelligence",Ergonomics
Eliminating combinatorics from the match in production systems (or rul
e-based systems) is important for expert systems, real-time performanc
e, machine learning (particularly with respect to the utility issue),
parallel implementations and cognitive modeling. In [74], the unique-a
ttribute representation was introduced to eliminate combinatorics from
the match. However, in so doing, unique-attributes engender a suffici
ently negative set of trade-offs, so that investigating whether there
are alternative representations that yield better trade-offs becomes o
f critical importance. This article identifies two promising spaces of
such alternatives, and explores a number of the alternatives within t
hese spaces. The first space is generated from local syntactic restric
tions on working memory. Within this space, unique-attributes is shown
to be the best alternative possible. The second space comes from rest
rictions on the search performed during the match of individual produc
tions (match-search). In particular, this space is derived from the co
mbination of a new, more relaxed, match formulation (instantiationless
match) and a set of restrictions derived from the constraint-satisfac
tion literature. Within this space, new alternatives are found that ou
tperform unique-attributes in some, but not yet all, domains.