• Refine Query
  • Source
  • Publication year
  • to
  • Language
  • 1
  • 1
  • Tagged with
  • 2
  • 2
  • 2
  • 2
  • 2
  • 2
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • About
  • The Global ETD Search service is a free service for researchers to find electronic theses and dissertations. This service is provided by the Networked Digital Library of Theses and Dissertations.
    Our metadata is collected from universities around the world. If you manage a university/consortium/country archive and want to be added, details can be found on the NDLTD website.
1

Algoritmos para o problema subset-sum em GPU

Vitor Venceslau Curtis 11 June 2013 (has links)
Este trabalho utiliza o problema subset-sum (SSP) como estudo de caso, com o objetivo de analisar a complexidade de paralelização em Unidades de Processamento Gráficas (GPU). O SSP foi escolhido por pertencer à classe dos problemas NP-Completo, possuir grande necessidade de memória e não ter cálculo de ponto flutuante, além de ser amplamente estudado na área acadêmica devido a sua importância prática e teórica. Estas características representam um desafio para paralelização em GPUs, pelo fato de serem especialistas em cálculos de ponto flutuante e por possuir pouca quantidade de memória em relação ao grande número de núcleos. Basicamente, são apresentados 3 novos algoritmos, implementados em linguagem CUDA C, com baixo consumo de memória: somente , onde , é a capacidade da mochila e é a quantidade de itens, ao invés de do paradigma de Bellman, referentes aos algoritmos do estado da arte implementados na mesma arquitetura. Esta característica permite um ganho significativo na quantidade de instâncias solucionáveis, além do melhor tempo computacional. Para uma variedade de benchmarks, obteve-se bons valores de speed-up em relação aos melhores resultados práticos conhecidos até agora. Isto foi possível graças a um novo método para a solução do SSP, permitindo sua computação em tempo e mesmo espaço, caso processadores sejam utilizados.
2

Improving the average time for solving subset-sum problem instances : algorithm variations and performance analysis

Davi Tassinari de Figueiredo 14 November 2013 (has links)
The Subset-sum Problem (SSP) consists of finding a subset of a set of integers whose sum is as close as possible to a target amount, without exceeding it. The problem is NP-complete, but there are dynamic programming-based algorithms which can solve many instances in reasonable time. Most of these algorithms attempt to minimize the time taken to find the solution for "hard" problem instances, but have sub-optimal performance when dealing with "easy" instances, either containing few items or with multiple equivalent solutions of which just one needs to be found. In this work, we analyze the characteristics of several well-known approaches, including standard dynamic-programming Bellman recursion, Horowitz-Sahni decomposition and core algorithms, and suggest variations which improve their average-case performance for easy instances, without harming their worst-case performance significantly. One variation improves the performance of the YS87 single state - multiple stages algorithm for sparse instances. Another minimizes the amount of work required by decomposition-based approaches before a solution is found, allowing the computation to be aborted earlier for dense instances with many exact solutions; this is achieved by using the programming concept of generators in order to compute partial solutions only when they are needed. We also examine the characteristics of several SSP test instance types which are commonly used to compare algorithm performances, analyzing how the distribution of partial solutions varies according to the type and size of the instances. The times taken by several algorithm implementations to solve each of the SSP instance types are measured and compared, and the variations in their behavior when dealing with each one are explained based on the instances'; characteristics.

Page generated in 0.1308 seconds