Abstract not found

In this paper we develop new Newton and conjugate gradient algorithms on the Grassmann and Stiefel manifolds. These manifolds represent the constraints that arise in such areas as the symmetric eigenvalue problem, nonlinear eigenvalue problems, electronic structures computations, and signal

Researchers m the areas of constramt sattsfactton problems, logtc programmmg, and truth mamtenance systems have suggested vartous schemes for enhancmg the performance of the backtracking algorithm Thts paper defines and compares the performance of three such schemes "backlumpmg," "learmng, " and "cycle-cutset " The backlumpmg and the cycle-cutset methods work best when the constramt graph is sparse, whde the learnmg scheme mostly benefits problem mstances wtth dense constramt graphs An integrated strategy ts described whwh utdtzes the dzstmct advantages of each scheme Expertments show that, m hard problems, the average tmprovement reahzed by the mtegrated scheme ts 20-25 % higher than any of the mdtvtdual schemes I.

Constraint logic programming languages such as CHIP [26,5] have demonstrated the practicality of declarative languages supporting consistency techniques and nondeterminism. Nevertheless they suffer from the black-box effect: the programmer must work with a monolithic, unmodifiable, inextensible

, to accommodate different types of constraints. Some constraints can be imposed without any modification of the algorithm, while others require the solution of a small associated linear system of equations. In particular, vertex location constraints, vertex normal constraints, and surface normal discontinuities

Constraint processing is complementary to logic programming: logic programming is concerned with proving theorems from a set # of sentences which are assumed to have a model; constraint processing is concerned with finding a model for #. This view of constraint processing has been realized

architecture for constraint processing with reactive agents. The resulting systems are very flexible and can be used to process dynamic and distributed problems. We define the local behaviors of the agents in this new approach and verify their collective behavior in the context of CSP.

First-order probabilistic models combine representational power of first-order logic with graphical models. There is an ongoing effort to design lifted inference algorithms for first-order probabilistic models. We analyze lifted inference from the perspective of constraint processing and, through

processing. The central processing mechmultiple levels simultaneously, including lexical, phonological, anism we invoke is the constraint satisfaction process that has morphological, syntactic, and text or discourse levels. At any been realized in interactive-activation models (e.g., Elman &

of three components: elements, form, and rationale. Elements are either processing, data, or connecting elements. Form is defined in terms of the properties of, and the relationships among, the elements-- that is, the constraints on the elements. The ratio-nale provides the underlying basis