Metabolomics
ISSN:2153-0769 JOM an open access journal
Conference Proceedings Open Access
Unay and Guzey, Metabolomics 2013, 3:2
DOI: 10.4172/2153-0769.1000120
Volume 3 • Issue 2 • 1000120
Particle swarm optimization
In the feld of computer science, PSO is a method that aims to
optimize a problem by trying to improve a possible solution with
iteration in limitation and manipulation of a pre-defned quality
measure. PSO processes an initial population of possible solutions,
dubbed particles in this case, and changing the position of these
particles in the search-space according to mathematical formula which
is consisting of the particles’ (a) position and (b) velocity. An individual
particle’s movement is altered according to its “local best known
position” and is also manipulated toward the “best known positions”
in the search-space. Te best known positions are updated as positions
which are more satisfactory for quality criteria, are discovered by other
particles. Tis mode of action is supposed to conduct the movement of
the swarm toward the best solutions [5].
PSO was frst intended for simulating social behavior, as a stylized
representation of the movement of organisms in a bird fock or fsh
school [6,7]. Te algorithm was simplifed and it was observed to be
performing optimization.
Particle swarm optimization on longest common subsequence
problem
Tis study uses PSO heuristic technique on LCSP. First, the
algorithm will take n sequences and generate an alphabet among all of
the distinct sequence elements without uncommon elements. Ten it
will generate a population of random sequences of the alphabet. Every
sequence will be a particle. It will do the evaluation with the technique,
occurrence evaluation, which will be described in detail in this paper.
Afer the evaluation, known local best score will be compared with the
global best score. If local best score is bigger, then it is the new global
best (initial global best is 0). Afer this, each particle move towards to
*Corresponding author: Meral Guzey, Department of Math and Life Sciences,
Main campus of University Maryland University College (UMUC), USA, E-mail:
meral.guzey@faculty.umuc.edu
Received June 25, 2013; Accepted August 06, 2013; Published August 13, 2013
Citation: Unay AT, Guzey M (2013) A Swarm Intelligence Heuristic Approach to
Longest Common Subsequence Problem for Arbitrary Number of Sequences.
Metabolomics 3: 120. doi:10.4172/2153-0769.1000120
Copyright: © 2013 Unay AT, et al. This is an open-access article distributed under
the terms of the Creative Commons Attribution License, which permits unrestricted
use, distribution, and reproduction in any medium, provided the original author and
source are credited.
A Swarm Intelligence Heuristic Approach to Longest Common
Subsequence Problem for Arbitrary Number of Sequences
Ali Teoman Unay
1
and Meral Guzey
2
1
Department of Intelligent Computing Systems, İzmir University of Economics, İzmir, Turkey
2
Department of Math and Life Sciences, Main campus of University Maryland University College (UMUC), USA
Introduction
In the clinical diagnosis of cancer, the corresponding biomarker
methods and measurements are based on the correct algorithms, which
are used for sequencing. Te necessity of development new or possibly
re-establishing previous biomarkers in the feld of cancer research
initiated us to work on new sequencing techniques.
Here, we describe “Longest Common Subsequence Problem”
(LCSP) with Particle Swarm Optimization (PSO), with the proposal of
novel “Occurrence Listing” (OL) technique as an evaluation function.
Previous studies shows, that despite the wide diferences between
popular approaches, like dynamic programming or other heuristic
methods, even there are some variations of dynamic programming
for three or more sequences [1], they generally work on two inputs
(sequences) [2,3]. Te benefts of our system are as follows: First, one
can work with minimum two or more sequences. Second, one has
fexibility of working with arbitrary number of sequences.
Longest common subsequence problem
LCSP dwells on longest common subsequence of two or more
sequences.
Although general case of a random number of input sequences, the
problem is NP-Hard, Dynamic programming can manage to solve the
problem on polynomial time provided that the number of is constant
[4].
What is “subsequence”?
A subsequence is a sequence that is created from another sequence
by excluding some elements but without changing the initial order. For
example, <A,B,D> sequence derived from <A,B,C,D,E,F> by omitting
element C, E and F.
Given two sequences X and Y, sequence G can be defned as a
common subsequence of X and Y, if G can be derived from both X and
Y individually.
For example,
if X=<A,B,C,D,E,G,C,E,D,B,G> and
Y=<B,E,G,C,F,E,U,B,K>
then a common subsequence of X and Y could be
G=<B,E,E>
Tis would not be the longest common subsequence. Te longest
common subsequence of X and Y is <B,E,G,C,E,B>.
Abstract
Personalized cancer care strategies involving sequencing requires accuracy. We aimed to develop a novel
approach to solve the longest common subsequence problem, which is a common computer science problem in the
feld of bioinformatics to facilitate the next generation sequencing of cancer biomarkers. We are using particle swarm
optimization heuristic technique, which uses a novel “Occurrence Listing” (OL) technique as the evaluation function. This
aims to keep lists of the sequence elements and offers criteria to evaluate randomly generated population of sequences.
M
e
t
a
b
o
l
o
m
i
c
s
:
O
p
e
n
A
c
c
e
s
s
ISSN: 2153-0769
Metabolomics: Open Access