Results 1  10
of
11
Performance Evaluation of Load Distribution Strategies in Parallel Branch and Bound Computations
 in Parallel Branch and Bound Computations Proc. 7th Symposium on Parallel and Distributed Processing (SPDP'95
, 1995
"... Load distribution is essential for efficient use of processors in parallel branchandbound computations because the computation generates and consumes nonuniform subproblems at runtime. This paper presents six decentralized load distribution strategies. They are incorporated in a runtime support s ..."
Abstract

Cited by 12 (1 self)
 Add to MetaCart
Load distribution is essential for efficient use of processors in parallel branchandbound computations because the computation generates and consumes nonuniform subproblems at runtime. This paper presents six decentralized load distribution strategies. They are incorporated in a runtime support system, and evaluated in the solution of set partitioning problems on two parallel computer systems. It is observed that local averaging strategies outperform the randomized allocation and the Acwn algorithm significantly in large scale system. They lead to an almost linear speedup in a PowerPCbased system with up to 32 nodes and to a speedup of 146.8 in a Transputerbased system with 256 nodes. It is also observed that the randomized allocation and the Acwn algorithm can be improved by 10% to 15% when the subproblem bound information is used in the decisionmaking. 1 Introduction Branchandbound is a wellknown technique for solving combinatorial search problems [4]. Its basic scheme is t...
A case study of load distribution in parallel view frustum culling and collision detection
 In Lecture Notes in Computer Science
, 2001
"... ..."
(Show Context)
Parallel Scalable Libraries and Algorithms for Computer Vision
, 1994
"... We describe a project that integrates applications requirements, parallel algorithm design, models of parallel computing, and software tools in order to improve the ability of applications researchers in the fields of computer vision and image processing (CVIP) to realize the performance potential o ..."
Abstract

Cited by 3 (1 self)
 Add to MetaCart
We describe a project that integrates applications requirements, parallel algorithm design, models of parallel computing, and software tools in order to improve the ability of applications researchers in the fields of computer vision and image processing (CVIP) to realize the performance potential of high performance parallel computers. This objective is achieved by pursuing four directions of research: the development of efficient and practical scalable algorithms for fundamental CVIP problems; the implementation of realistic CVIP scenarios on high performance computers; the development of a scalable, architectureindependent parallel model and a metric for estimating the performance of an algorithm on an existing parallel machine; and the development of prototype software tools. In this paper we describe work in each of these areas. 1 Introduction The goal of this project is to develop both theory and tools to make scalable high performance parallel computing more readily available ...
Parallel and Distributed BranchandBound/A* Algorithms
, 1994
"... In this report, we propose new concurrent data structures and load balancing strategies for BranchandBound (B&B)/A* algorithms in two models of parallel programming : shared and distributed memory. For the shared memory model (SMM), we present a general methodology which allows concurrent man ..."
Abstract

Cited by 2 (0 self)
 Add to MetaCart
In this report, we propose new concurrent data structures and load balancing strategies for BranchandBound (B&B)/A* algorithms in two models of parallel programming : shared and distributed memory. For the shared memory model (SMM), we present a general methodology which allows concurrent manipulations for most tree data structures, and show its usefulness for implementation on multiprocessors with global shared memory. Some priority queues which are suited for basic operations performed by B&B algorithms are described : the Skewheaps, the funnels and the Splaytrees. We also detail a specific data structure, called treap and designed for A* algorithm. These data structures are implemented on a parallel machine with shared memory : KSR1. For the distributed memory model (DMM), we show that the use of partial cost in the B&B algorithms is not enough to balance nodes between the local queues. Thus, we introduce another notion of priority, called potentiality, between nodes that take...
P2P B&B and GA for the FlowShop Scheduling Problem
, 2008
"... Solving exactly Combinatorial Optimization Problems (COPs) using a BranchandBound algorithm (B&B) requires a huge amount of computational resources. The efficiency of such algorithm can be improved by its hybridization with metaheuristics such as Genetic Algorithms (GA) which proved their ef ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
Solving exactly Combinatorial Optimization Problems (COPs) using a BranchandBound algorithm (B&B) requires a huge amount of computational resources. The efficiency of such algorithm can be improved by its hybridization with metaheuristics such as Genetic Algorithms (GA) which proved their effectiveness, since they generate acceptable solutions in a reasonable time. Moreover, distributing at large scale the computation, using for instance PeertoPeer (P2P) Computing, provides an efficient way to reach high computing performance. In this chapter, we propose ParallelBB and ParallelGA, which are P2Pbased parallelization of the B&B and GA algorithms for the computational Grid. The two algorithms have been implemented using the ProActive distributed object Grid middleware. The algorithms have been applied to a monocriterion permutation flowshop scheduling problem and promisingly experimented on the Grid5000 computational Grid.
Reporting Computational Experiments . . .
 ORSA JOURNAL ON COMPUTING
, 1992
"... Accompanying the increasing availability of parallel computing technology is a corresponding growth of research into the development, implementation, and testing of parallel algorithms. This paper examines issues involved in reporting on the empirical testing of parallel mathematical programming alg ..."
Abstract
 Add to MetaCart
Accompanying the increasing availability of parallel computing technology is a corresponding growth of research into the development, implementation, and testing of parallel algorithms. This paper examines issues involved in reporting on the empirical testing of parallel mathematical programming algorithms, both optimizing and heuristic. We examine the appropriateness of various performance metrics and explore the effects of testing variability, machine influences, testing biases, and the effects of tuning parameters. Some of these difficulties were explored further in a survey sent to leading computational mathematical programming researchers for their reactions and suggestions. A summary of the survey and proposals for conscientious reporting are presented.
Parallel Scalable Libraries and Algorithms for Computer Vision*
"... We describe a project that integrates applications requirements, parallel algorithm design, models of parallel computing, and software tools in order to improve the ability of applications researchers in the fields of computer vision and image processing (CVIP) to realize the performance potential ..."
Abstract
 Add to MetaCart
We describe a project that integrates applications requirements, parallel algorithm design, models of parallel computing, and software tools in order to improve the ability of applications researchers in the fields of computer vision and image processing (CVIP) to realize the performance potential of high performance parallel computers. This objective is achieved by pursuing four directions of research: the development of e@ cient and practical scalable algorithms for fundamental CVIP problems; the implementation of realistic CVIP scenarios on high performance computers; the development of a scalable, architectureindependent parallel model and a metric for estimating the performance of an algorithm on an existing parallel machine; and the development of prototype software tools. In this paper we describe work in each of these areas. 1
Developing Scheduling Algorithms to Access the Critical Section in SharedMemory Parallel Computers
, 1997
"... ree of interactions among the concurrent processes. A sharedmemory multiprocessor machine consists of two main parts: a set of independent processors and a global shared memory. The global memory contains a set of modules connected to and shared by all processors through a shared highspeed bus. ..."
Abstract
 Add to MetaCart
ree of interactions among the concurrent processes. A sharedmemory multiprocessor machine consists of two main parts: a set of independent processors and a global shared memory. The global memory contains a set of modules connected to and shared by all processors through a shared highspeed bus. In sharedmemory multiprocessor computer, we face a critical resource constraint, namely, the shared memory. This resource constraint can inhibit the efficient cooperation of the parallel processors to exchange the partial results needed to solve a problem. In such environment, shared variables are used to facilitate communication between concurrent processes. But these shared variables must be protected from nondeterminism, which can result from concurrent access by more than one process at a time. These shared variables can be protected from nondeterminism by placing the code that uses them in a critical section that can be executed by only one process at a time. Unfortunately, accessing t
Parallelization Of Irregular Algorithms For Shape Detection
 International Conference on Image Processing
, 1996
"... There are a series of very efficient sequential algorithms that generate irregular trees during the process of detecting shapes in images. These algorithms are based on the Fast Hough Transform and are used for solving the most complex stages of detection when the production of the parameters is unc ..."
Abstract
 Add to MetaCart
(Show Context)
There are a series of very efficient sequential algorithms that generate irregular trees during the process of detecting shapes in images. These algorithms are based on the Fast Hough Transform and are used for solving the most complex stages of detection when the production of the parameters is uncoupled. However, the parallelization of these algorithms is complex, and the problem of load distribution is crucial. In this work we present three parallel algorithms for solving this problem. One of the solutions employs static load balancing. The other two use dynamic balancing with two different control policies: distributed and centralized. These algorithms may also be used for solving other problems, such as the Branch and Bound, that generate irregular trees. 1. INTRODUCTION The Hough Transform (HT) is a method that has been shown to be very efficient for the detection of shapes in binary images. However, it presents problems having to do with its computational complexity that have l...
A Probabilistic Model for BestFirst Search BB Algorithms
, 1997
"... . In this paper we present a probabilistic model for analysing trees generated during problem solving using bestfirst search Branch and Bound algorithm. We consider the weight associated with the edges as well as the number of children in a node to be random. The expressions obtained for the sequen ..."
Abstract
 Add to MetaCart
. In this paper we present a probabilistic model for analysing trees generated during problem solving using bestfirst search Branch and Bound algorithm. We consider the weight associated with the edges as well as the number of children in a node to be random. The expressions obtained for the sequential algorithm are applied to several parallel strategies using a multiprocessor with distributed memory. 1 Introduction Branch and Bound (BB) algorithms are the most frequently used methods for the solution of combinatorial optimization problems. There are four fundamental ingredients for a BB method: branching rule, bounding procedure, selection rule and termination test. The branching rule takes a given problem G and either solves it directly or derives from it a set of subproblems G 1 , G 2 , ..., G a (expansion of G). The bounding procedure computes a lower bound on the cost of an optimal solution to a subproblem and eliminates the subproblems that can not generate an optimal solution...