PDC Review 1

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 6

CSE 4001 PARALLEL AND DISTRIBUTED COMPUTING

REVIEW- 1(J COMPONENT)

TITLE: PARALLEL EXECUTION OF 0/1KNAPSACK PROBLEM


(STUDENT MAXIMUM MARKS EVALUATION)

GROUP MEMBERS:
1. S.SRI HARSHA-18BCE0401
2. A.SHANKAR SUDHEER-18BCE0556

SLOT: C2
FACULTY: PROF.NARAYANAMOOTHI.M
INTRODUCTION

1. OBJECTIVE

The objective of this is to solve the knapsack problem using parallelism and to
use this algorithm to in Student marks evaluation system in considering the
right choices and to reduce the amount of time take to it. This project looks at
existing and common approaches to solving the knapsack problem like
dynamic programming suggest a parallel approach to solving the problem
utilizing multiple cores of a computer using threads to get a solution in
optimum time.

2. MOTIVATION
The problem is if a question paper has choices and If each question weightage
is in heterogeneous pattern (all questions are not of equal marks). Then a
knapsack algorithm must be used to calculate the marks scored by students
If the college student count is more then this takes a lot of time.

3. ABSTRACT

The main aim of our project is to show the benefits of parallelization by a


direct comparison of the time taken in common approaches used to solve the
0/1 knapsack problem, to that of using parallelism to solve it and view the
benefits and drawbacks that are present in taking this approach.We also
present the constraints involved in this approach and also puts forward the
OpenMP optimal programming model which is used to this application
4. LITERATURE SURVEY

(i) A parallel algorithm for the knapsack problem using a generation


and searching technique.
-Henry Ker Chang Chag, Jonathan Jen Rong Chen, Snyong-Jian-Shyu The
purpose of this paper was to propose a new parallel algorithm for the
knapsack problem. The authors developed a generation and searching
technique to derive the derived two ordered tests in the preliminary process of
general knapsack problem. The proposed parallel algorithm was developed on
the basis of SMID machine with shared memory. There algorithm needed
0(2*nu) memory and 0(2*n) processors.

The given paper gives the content of how parallelization can be applied to
solve the category of combined problem -0/1 knapsack which are
complimenting in our project.

(ii) Computing Partitions with Applications to the Knapsack Problem.


-Horowitz, Ellis, and Sartaj Sahni.
In this paper authors have investigated on all the usual algorithms for this
problem in terms of both asymmetric computing time and storage
requirements as well as average computing time.
They developed a technique which improved all of the dynamic programming
methods by a square root factor. A new branch and bound algorithm which is
significantly faster than the green tag and Hegerich algorithm was also
presented.
Since we are also comparing technique before going for the parallel solution,
this paper provided the approach for the comparing of algorithm keeping
knapsack problem in contest.

(iii) Optimal parallel algorithms for the knapsack problems without


memory conflicts.-Li, Ken-Li, Ren-Fa Li, and Qing-Hua Li.
The authors discussed importance of the knapsack problem in the application
of crypto system and in theory with the imphasis to find the techniques that
could lead to practical algorithms with less run time. The paper prosed a new
parallel algorithm for the knapsack problem where the merging algorithm is
adopted.
While reviewing various techniques for parallelizing the algorithm using this
paper context we can also optimize the algorithm for memory efficiency
using openMP.

(iv) OPENMP: An Industry-Standard API for shared memory


programming The author present a new way to achieve scalability in parallel
software with OpenMP, their portable alternative to message passing. They
discuss its capabilities through specific examples and comparisons with other
standard parallel programming model. Since we are implementing the
parallelization of the 0/1 knapsack problem this paper provide necessary
content to ways of applying OpenMP to increase parallelization, efficiency in
your code with use of appropriate drivers

(v) Concurrency and Parallelism


-SASH GOLOSHIFIN, DIMA ZORBALEV, JOO FLATOW
The authors discussed the exponential increase in the processing power of
computer systems with the processors getting faster with every model and
programs designed to challenge the hardware resources of an expensive
workstation. They also discuss the fact that the no of processor are observing
exponential increase in contrast to increase in there speed. The paper also
presented a healthy discussion with memory and speed argument.

The paper discussed various practices which could help in implementing


parallelization in our project and also presented theoretical basis for arguing
that parallel execution could led to better implementation of knapsack
problem.

You might also like