QUANTITATIVE ANALYSIS OF TABU SEARCH ALGORITHM FOR A VLSI PLACEMENT APPLICATION

QUANTITATIVE ANALYSIS OF TABU SEARCH ALGORITHM FOR A VLSI PLACEMENT APPLICATION

Bibliographic Details
Main Author: SHARMA, VIKAS
Language:English
Published: University of Cincinnati / OhioLINK 2001
Subjects:
Tabu
tabu tenure
tenure
candidate list
Tabu Search
Wirelength
Algorithm
Online Access:http://rave.ohiolink.edu/etdc/view?acc_num=ucin990820742
id ndltd-OhioLink-oai-etd.ohiolink.edu-ucin990820742
record_format oai_dc
spelling ndltd-OhioLink-oai-etd.ohiolink.edu-ucin9908207422021-08-03T06:16:16Z QUANTITATIVE ANALYSIS OF TABU SEARCH ALGORITHM FOR A VLSI PLACEMENT APPLICATION SHARMA, VIKAS Tabu tabu tenure tenure candidate list Tabu Search Wirelength Algorithm The abundance of difficult combinatorial optimization problems in practical settings such as telecommunications, financial planning and VLSI design automation has motivated the development of powerful optimization techniques. Simulated Annealing, Genetic algorithms, Network Flow and other heuristics have long been explored for solvingthese problems. During the recent years, miniaturization of the chip sizes and increasing densities of logic have posed a continuous design-related challenge to the VLSI design automation community. Promising results have been reported from the application of Tabu Search algorithm to VLSI circuit partitioning, floorplanning, placement and routing. Tabu Search leverages the ability to store solutions already visited and to make strategic solutions on the basis of the stored information for achieving optimal solutions. This thesis presents a quantitative analysis of the various parameters of Tabu Search. In particular, we look at the techniques of moving out of the current search space called diversification methodologies and strategies for determining the best solution in the current search space called intensification methodologies. This is achieved by applying a tabu tenure on the moves, thus forbidding execution of certain moves for some iterations. In addition, we generate a candidate list of moves for selecting moves for future executions. Our approach makes use of a Tabu Search-based Force Directed placement technique. We perform physical placements of VLSI circuits on a two-dimensional array. In the first step, the circuit cells are placed randomly and their respective tendencies to move in all the four directions are computed. This is followed by selection of the moves for swap on the basis of candidate list implementations. Intensification and diversification phases are repeated alternately. An optimized ratio of the number of iterations of intensification and diversification phases ensures good quality solutions with low processing time. We have demonstrated that our Tabu Search-based placement approach achieves a speed up ranging from 5 to 17 times over a similarly implemented Simulated Annealing-based placement algorithm, depending on the circuit size. 2001-10-11 English text University of Cincinnati / OhioLINK http://rave.ohiolink.edu/etdc/view?acc_num=ucin990820742 http://rave.ohiolink.edu/etdc/view?acc_num=ucin990820742 unrestricted This thesis or dissertation is protected by copyright: all rights reserved. It may not be copied or redistributed beyond the terms of applicable copyright laws.
collection NDLTD
language English
sources NDLTD
topic Tabu
tabu tenure
tenure
candidate list
Tabu Search
Wirelength
Algorithm
spellingShingle Tabu
tabu tenure
tenure
candidate list
Tabu Search
Wirelength
Algorithm
SHARMA, VIKAS
QUANTITATIVE ANALYSIS OF TABU SEARCH ALGORITHM FOR A VLSI PLACEMENT APPLICATION
author SHARMA, VIKAS
author_facet SHARMA, VIKAS
author_sort SHARMA, VIKAS
title QUANTITATIVE ANALYSIS OF TABU SEARCH ALGORITHM FOR A VLSI PLACEMENT APPLICATION
title_short QUANTITATIVE ANALYSIS OF TABU SEARCH ALGORITHM FOR A VLSI PLACEMENT APPLICATION
title_full QUANTITATIVE ANALYSIS OF TABU SEARCH ALGORITHM FOR A VLSI PLACEMENT APPLICATION
title_fullStr QUANTITATIVE ANALYSIS OF TABU SEARCH ALGORITHM FOR A VLSI PLACEMENT APPLICATION
title_full_unstemmed QUANTITATIVE ANALYSIS OF TABU SEARCH ALGORITHM FOR A VLSI PLACEMENT APPLICATION
title_sort quantitative analysis of tabu search algorithm for a vlsi placement application
publisher University of Cincinnati / OhioLINK
publishDate 2001
url http://rave.ohiolink.edu/etdc/view?acc_num=ucin990820742
work_keys_str_mv AT sharmavikas quantitativeanalysisoftabusearchalgorithmforavlsiplacementapplication
_version_ 1719433801443049472

Similar Items