Abstract—A novel Fast Bacterial Swarming Algorithm
(FBSA) for high-dimensional funct ion optimization is presented
in this paper. The proposed algorithm combines the foraging
mechanism of E-coli bacteriumintroducedinBacterial
Foraging Algorithm (BFA) with the swarming pattern of birds
in block introduced in Particle Swarm Optimization (PSO). It
incorporates the merits of the two bio-inspired algorithms to
improve the convergence for high-dimensional function
optimization. A new parameter called attraction factor is
introduced to adjust the bacterial trajectory according to the
location of the best bacterium (bacterium w ith best fitness value).
An adaptive step length is adopted to improve the local search
ability. The algorithm has been evaluated on standard
high-dimensional benchmark functions in comparison with
BFA and PSO respectively. The simulation results have
demonstrated the fast convergence ability and the improved
optimization accuracy of FBSA.
I. INTRODUCTION
ANY bio-inspired computational methodologies such
as Genetic Algorithm (GA) [1], Ant Colony
Optimization (ACO) [2], Particle Swarm Optimization
(PSO) [3] and Artificial Fish Swarm Algorithm (AFSA) [4]
have been intensively studied and applied to various
optimization problems. In r ecent years, a new and rapidly
growing subject - Bacterial Foraging Algorithm (BFA),
which is inspired by the behavior that E-coli bacteria
searching for food in human intestines, attracts more and
more attentions and shows its advantage in global
optimization. BFA has been applied to many kinds of real
world optimization problems, such as adaptive control [5],
harmonic signal estimation [6], optimal power system
stabilizers design [7] and optimal power flow [8][9].
Although BFA has prominent and encouraging performance
for global optimization problems as reported in [5]-[9], our
simulation results had proved th at the algorithm is
time-consuming. A key advance will therefore be met by a
significant reduction in computational time-costs whilst
further improving the global search capabilities of the
algorithm [10].
In BFA, eac h bacterium releases attractants to signal other
This work wa s supported by the National Natural Science Foundation of
China under Grant 60572100 and by the R oyal Society (U.K.) International
Joint Projects 2006/R3 - Cost Share with NSFC under Grant 60711130233.
Ying Chu, Hua Mi, Huilian Liao and Zhen Ji ar e with the Texas
Instruments DSPs Lab, Shenzhen University, Shenzhen, 518060, China
(e-mail: chuying@szu.edu.cn; mihua1984@163.com; liaohuilian@163.com;
phone: 86-755-26557413; fax: 86-755-26536198; e-mail:
jizhen@szu.edu.cn).
Q. H. Wu is with the Department of Electrical Engineering and
Electronics, The University of Liverpool, Liverpool, L69 3GJ, UK (e-mail:
qhwu@liv.ac.uk).
bacteria to swarm together and to keep a safe distance with
each other. Such a mechanism to some extent suppresses the
convergence because the bacteria in nutrient-poo r areas
attract the bacteria in nutrient-ric h areas, which slow s dow n
the convergence speed of the entire group. The repelling
effect among bacteria also prevents the group gathering
together. Inspired by the swarming pattern in PSO [11][12], a
novel principle of swarming is introduced into Fast Bacterial
Swarming Algorithm (FBSA) for the first time to reduce the
computation load of the algorithm. Each bacterium adjusts its
position according to the n eighborh ood environment and the
position of the best bacterium alternatively. Simulation
results have shown that the presented swarming mechanism
does provide faster and more efficient convergence. An
adaptive step length is also introduced to improve the
precision of the optimization results. It h as been proved by
simulations that the proposed adjustment effectively prevents
the bacteria being trapped into local optima. In order to
demonstrate the merits of th e proposed FBSA, thirteen
mathematical benc hmark functions [13], whic h cover
high-dimensional unimodal and multimodal function s, were
selected to evaluate the algorithm. Simulation results have
shown that FBSA perfo rms better and converges much
rapidly in comparison with B FA and PSO fo r
high-dimensional function optimization.
This paper is organized as follows: Section II describes the
basic principle of BFA. Section III introduces the proposed
FBSA. The simulation results and discussion s are shown in
Section IV, followed by conclusions in Section V.
II. B
ACTERIAL FORAG ING ALGORITHM
The optimization in BFA comprises the followin g
proc e sses: chemotaxis, swarming, reproduction, elimination
and dispersal. Chemotaxis is the activity that bacteria
gathering to nutrient-rich areas spontaneously. A cell-to-cell
communication mechanism is established to simulate the
biological behavior of bacteria swarming. Reproduction
comes from the concept of natural selection and only the
bacteria best adapted to their environment tend to survive and
trans mit their genetic characters to succeeding ge nerations
while those less adapted tend to be eliminated.
Elimination-dispersal event selects parts of the bacteria to
diminish and disperse into random position s in the
environment, which ensures the diversity of th e species.
A. Chemotaxis
An E-coli bacterium can move in two different ways
alternatively: tumble and run.Atumble is represented by a
unit walk with random direction, a unit walk with the same
direction as the previous step indicates a run. A chemotactic
A Fast Bacterial Swarming Algorithm For High-dimensional
Function Optimization
Yin
g Chu, Hua Mi, Huilian Liao, Zhe
Ji, and Q. H. Wu
M
3135
978-1-4244-1823-7/08/$25.00
c
2008 IEEE
Authorized licensed use limited to: SUN YAT-SEN UNIVERSITY. Downloaded on April 5, 2009 at 08:31 from IEEE Xplore. Restrictions apply.