Results 1 
7 of
7
Convex Sets Of Probabilities Propagation By Simulated Annealing
 In Proceedings of the Fith International Conference IPMU'94
, 1994
"... An approximated simulation algorithm is presented for the propagation of convex sets of probabilities. It is assumed that the graph is such that an exact probabilistic propagation is feasible. The algorithm is a simulated annealing procedure, which randomly selects probability distributions among th ..."
Abstract

Cited by 22 (5 self)
 Add to MetaCart
An approximated simulation algorithm is presented for the propagation of convex sets of probabilities. It is assumed that the graph is such that an exact probabilistic propagation is feasible. The algorithm is a simulated annealing procedure, which randomly selects probability distributions among the possible ones, performing at the same time an exact probabilistic propagation. The algorithm can be applied to general directed acyclic graphs and is carried out on a tree of cliques. Some experimental tests are shown. 1. Introduction One of the main problems with probabilistic propagation algoritms on graphical structures is the introduction of the initial exact conditional probabilities. A number of authors have tried to overcome this difficulty by allowing the use of intervals on the specified probabilities [5, 10, 11, 3, 14, 13, 19, 23, 2]. Some of these works [10, 11, 3, 13, 23] focus on the use of convex sets of probabilities. Convex sets are a more general tool for representing un...
A New Evolutionary Approach to the DegreeConstrained Minimum Spanning Tree Problem
 IEEE Transactions on Evolutionary Computation
, 1999
"... Finding the degreeconstrained minimum spanning tree (dMST) of a graph is a wellstudied NPhard problem of importance in communications network design and other networkrelated problems. In this paper we describe some previously proposed algorithms for solving the problem, and then introduce a nove ..."
Abstract

Cited by 10 (2 self)
 Add to MetaCart
Finding the degreeconstrained minimum spanning tree (dMST) of a graph is a wellstudied NPhard problem of importance in communications network design and other networkrelated problems. In this paper we describe some previously proposed algorithms for solving the problem, and then introduce a novel tree construction algorithm called the Randomised Primal Method (RPM) which builds degreeconstrained trees of low cost from solution vectors taken as input. RPM is applied in three stochastic iterative search methods: simulated annealing, multistart hillclimbing, and a genetic algorithm. While other researchers have mainly concentrated on finding spanning trees in Euclidean graphs, we consider the more general case of random graph problems. We describe two random graph generators which produce particularly challenging dMST problems. On these and other problems we find that the genetic algorithm employing RPM outperforms simulated annealing and multistart hillclimbing. Our experimental ...
Discovering planar segregations
 Proceedings of AAAI Spring Symposium on Machine Learning of Natural Language and Ontology
, 1991
"... this report I present an algorithm for finding planar segregations of phonemes for particular languages. This algorithm requires no domainspecific knowledge of phonology or phonetics. Despite this lack of knowledge, the implemented algorithm has identified the structurally significant segregations ..."
Abstract

Cited by 5 (1 self)
 Add to MetaCart
this report I present an algorithm for finding planar segregations of phonemes for particular languages. This algorithm requires no domainspecific knowledge of phonology or phonetics. Despite this lack of knowledge, the implemented algorithm has identified the structurally significant segregations for thirty languages
Large deviations for a class nonhomogeneous Markov chains
, 2005
"... Large deviation results are given for a class of “perturbed ” nonhomogeneous Markov chains on finite state space which formally includes some stochastic optimization algorithms. Specifically, let {Pn} be a sequence of transition matrices on a finite state space which converge to a limit transition m ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
Large deviation results are given for a class of “perturbed ” nonhomogeneous Markov chains on finite state space which formally includes some stochastic optimization algorithms. Specifically, let {Pn} be a sequence of transition matrices on a finite state space which converge to a limit transition matrix P. Let {Xn} be the associated nonhomogeneous Markov chain where Pn controls movement from time n − 1 to n. The main statements are a large deviation principle and bounds for additive functionals of the nonhomogeneous process under some regularity conditions. In particular, when P is reducible, three regimes depending on the decay of certain “connection ” Pnprobabilities are identified. Roughly, if the decay is too slow, too fast, or in an intermediate range, the large deviation behavior is trivial, same as the timehomogeneous chain run with P, or nontrivial and involving the decay rates. Examples of anomalous behaviors are also given when the approach Pn → P is irregular. Results in the intermediate regime apply to “geometrically fast ” running optimizations, and to some issues in glassy physics. Research supported in part by NSF grant NSF/DMS0003811. Key words and phrases: large deviations, nonhomogeneous, Markov, optimization,
Invited Review Fuzzy scheduling: Modelling flexible constraints vs. coping with incomplete knowledge
"... An overview of some fuzzy setbased approaches to scheduling is proposed,emphasizing two distinct uses of fuzzy sets: representing preference profiles and modelling uncertainty distributions. The first setting leads to a valued,noncompensatory generalization of constraintdirected scheduling. The ot ..."
Abstract
 Add to MetaCart
An overview of some fuzzy setbased approaches to scheduling is proposed,emphasizing two distinct uses of fuzzy sets: representing preference profiles and modelling uncertainty distributions. The first setting leads to a valued,noncompensatory generalization of constraintdirected scheduling. The other setting yields a possibilitytheoretic counterpart of PERT,where probability distributions of activity durations are changed into possibility distributions,for the purpose of modelling incomplete information. It is pointed out that a special case of the latter,intervalvalued PERT,is a difficult,illknown problem,regarding the determination of critical activities,latest starting times and floats. Lastly when flexible constraints and uncertain processing times are to be jointly considered,the use of possibilistic decision theory leads to the computation of robust schedules.
• City of GrazA BALANCED ACCUMULATION SCHEME FOR PARALLEL PDE SOLVERS
, 2013
"... Abstract. We present a load balancing technique for a boundary data accumulation algorithm for nonoverlapping domain decompositions. The technique is used to speed up a parallel conjugate gradient algorithm with an algebraic multigrid preconditioner to solve a potential problem on an unstructured t ..."
Abstract
 Add to MetaCart
Abstract. We present a load balancing technique for a boundary data accumulation algorithm for nonoverlapping domain decompositions. The technique is used to speed up a parallel conjugate gradient algorithm with an algebraic multigrid preconditioner to solve a potential problem on an unstructured tetrahedral finite element mesh. The optimized accumulation algorithm significantly improves the performance of the parallel solver and we show a nearly 50 percent runtime improvement over the standard approach in a benchmark run with 48 MPI processes. The load balancing problem itself is a global optimization problem that is solved approximately by local optimization algorithms in parallel that require no communication during the optimization process. 1.
A General Parallel Tabu Search Algorithm For Combinatorial Optimisation Problems?
"... Abstract. Tabu Search (TS) is a metaheuristic search algorithm that is easy to parallelise. E cient parallelisation of TS can represent a signicantsaving in the realtime required to solve a problem over an equivalent sequential algorithm. In this study, a general parallel TS algorithm for solving ..."
Abstract
 Add to MetaCart
Abstract. Tabu Search (TS) is a metaheuristic search algorithm that is easy to parallelise. E cient parallelisation of TS can represent a signicantsaving in the realtime required to solve a problem over an equivalent sequential algorithm. In this study, a general parallel TS algorithm for solving combinatorial optimisation problems (COPs) is presented. The unique feature of our approach is that the TS solves a wide range of COPs expressed in a high level syntax. The bene t of this general code is that it can be used in realtime applications due to its parallel scalability and the fact that it can accept changing problem de nitions. After reviewing a number of suitable parallelisation strategies, results are presented that show that good parallel speedup is achieved while e cient solutions to hard COPs are obtained.