INSTITUTE OF TECHNOLOGY, NIRMA UNIVERSITY, AHMEDABAD – 382 481, 08-10 DECEMBER, 2011 1

Bacterial foraging optimization algorithm: A

Derivative free technique
Nitin Kumar Jhankal, Dipak Adhyaru, Member, IEEE

more likely to enjoy reproductive success (they obtain

Abstract--Bacterial foraging optimization algorithm (BFOA) enough food to enable them to reproduce). After many
has been widely accepted as a global optimization algorithm of generations, poor foraging strategies are either eliminated or
current interest for optimization and control. BFOA is inspired
shaped into good ones (redesigned). Logically, such
by the social foraging behaviour of Escherichia coli. BFOA has
already drawn the attention of researchers because of its evolutionary principles have led scientists in the field of
efficiency in solving real-world optimization problems arising ”foraging theory” to hypothesize that it is appropriate to
in several application domains. In present paper, a detailed model the activity of foraging as an optimization process: A
explanation of this algorithm is given. Comparative analysis of foraging animal takes actions to maximize the energy
BFOA with Genetic Algorithm (GA) is presented. obtained per unit time spent foraging.
Index Terms--About four, alphabetical order, key words or
phrases, separated by commas.
II. BASIC DETAILS: BACTERIA FORAGING TECHNOLOGY
I. INTRODUCTION
A. Foraging: Element of Foraging Theory

O ptimization problems defined by functions for which

derivatives are unavailable or available at a prohibitive
Foraging theory is based on the assumption that animals
search for and obtain nutrients in a way that maximizes their
cost are appearing more and more frequently in energy intake E per unit time T spent foraging. Hence, they
computational science and engineering. Increasing try to maximize a function like E/T (or they maximize their
complexity in mathematical modelling, higher sophistication long-term average rate of energy intake). Maximization of
of scientific computing, and abundance of legacy codes are such a function provides nutrient sources to survive and
some of the reasons why derivative-free optimization is additional time for other important activities (e.g., fighting,
currently an area of great demand. fleeing, mating, reproducing, sleeping, or shelter building).
In many physical applications, the true models or Optimal foraging theory formulates the foraging
functions being optimized are extremely expensive to problem as an optimization problem and via computational
evaluate but, based e.g. on simplified physics or mesh or analytical methods [2].
coarsening, there are often surrogate models available, less B. Search Strategies for Foraging
accurate but cheaper to evaluate. In these circumstances, one Some animals are “cruise” or “ambush” searchers. For
would expect to design an optimization framework capable the cruise approach to searching, the forager moves
of extracting as much information as possible from the continuously through the environment, constantly searching
surrogate model while parsimoniously using the fine, true for prey at the boundary of the volume being searched (tuna
model to accurately guide the course of the optimization fish and hawks are cruise searchers). In ambush search, the
process.[1] forager (e.g., a rattlesnake) remains stationary and waits for
prey to cross into its strike range [2]. The search strategies of
many species are actually between the cruise and ambush
A. Different Types of Derivative free Optimization
extremes.
1) Genetic Algorithm
2) Simulated Analysis C. Bacterial Foraging: E.coli
3) Random Search Method The E. coli bacterium has a plasma membrane, cell wall,
4) Swarm Optimization and capsule that contains the cytoplasm and nucleoid (Figure
5) Ant Colony Algorithm 1). The pili (singular, pilus) are used for a type of gene
6) Bacterial Foraging transfer to other E. coli bacteria, and flagella (singular,
Natural selection tends to eliminate animals with poor flagellum) are used for locomotion. The cell is about 1 μmin
“foraging strategies” (methods for locating, handling, and diameter and 2 μm in length [1]. The E. coli cell only
ingesting food) and favor the propagation of genes of those weighs about 1 picogram and is about 70% water.
animals that have successful forging strategies since they are Salmonella typhimurium is a similar type of bacterium.

978-1-4577-2168-7/11/$26.00 ©2011 IEEE

2 INTERNATIONAL CONFERENCE ON CURRENT TRENDS IN TECHNOLOGY, „NUiCONE – 2011‟

D. Swimming and Tumbling via Flagella k, l) represents the ith bacterium at jth chemotactic, kth
Locomotion is achieved via a set of relatively rigid reproductive, and lth elimination–dispersal step. C (i) is a
flagella that enable the bacterium to swim via each of them scalar and indicates the size of the step taken in the random
rotating in the same direction at about 100-200 revolutions direction specified by the tumble (run length unit). Then, in
per second (in control systems terms, we think of the computational chemotaxis, the movement of the bacterium
flagellum as providing for actuation). Each flagellum is a may be represented by
left-handed helix configured so that as the base of the
flagellum (i.e., where it is connected to the cell) rotates i i (i)
counterclockwise, as viewed from the free end of the
( j 1, k , l ) ( j, k , l ) C (i) (1)
T
flagellum looking toward the cell, it produces a force against
(i) (i)
the bacterium so it pushes the cell. where Δ indicates a unit length vector in the random
direction.
2) Swarming:
Interesting group behaviour has been observed for several
motile species of bacteria including E.coli and S.
typhimurium, where stable spatiotemporal patterns (swarms)
are formed in semisolid nutrient medium. A group of E.coli
cells arrange themselves in a traveling ring by moving up the
nutrient gradient when placed amid a semisolid matrix with
a single nutrient chemo-effecter. The cells when stimulated
by a high level of succinate release an attractant aspartate,
which helps them to aggregate into groups and, thus, move
as concentric patterns of swarms with high bacterial density.
3) Reproduction:
The least healthy bacteria eventually die while each of
the healthier bacteria (those yielding lower value of the
Fig. 1. Bacterial foraging E.coli [7]
objective function) asexually split into two bacteria, which
An E. coli bacterium can move in two different ways; If are then placed in the same location. This keeps the swarm
the flagella rotate clockwise, each flagellum pulls on the cell, size constant.
and the net effect is that each flagellum operates relatively 4) Elimination and Dispersal:
independently of the others, and so the bacterium “tumbles” To simulate this phenomenon in BFOA, some bacteria
about (i.e., the bacterium does not have a set direction of are liquidated at random with a very small probability while
movement and there is little displacement See Fig. 2. (a) [4]) the new replacements are randomly initialized over the
If the flagella move counterclockwise, their effects search space.
accumulate by forming a bundle (it is thought that the
bundle is formed due to viscous drag of the medium), and IV. EXAMPLE & SIMULATION
hence they essentially make a composite propeller and push
the bacterium so that it runs (swims) in one direction (see A. Function Optimization via Bacterial Foraging
Fig. 2 (a) [5]). As a simple illustrative example [2], we use the
algorithm to try to Find minimum of function in Figure 4
([15, 5] is the global minimum point, [20, 15] is a local
minimum). Standard ideas from optimization theory can be
used to set the algorithm parameters.

Fig. 2. Swimming, tumbling and chemotactic behavior of E coli

III. BACTERIAL FORAGING OPTIMIZATION ALGORITHM

A. Steps For BFOA

1) Chemotaxis:
This process simulates the movement of an E.coli cell
through swimming and tumbling via flagella. Suppose θi(j,
INSTITUTE OF TECHNOLOGY, NIRMA UNIVERSITY, AHMEDABAD – 382 481, 08-10 DECEMBER, 2011 3

Fig. 3. Function with multiple extremes point

B. Simulation experiment: Fig. 4. BFOA resuls

A function to be optimized is created with following

MATLAB program:
function fposition=Live_fn(x)
p=0;q=0;
for k=1:5
p=p+k*cos((k+1)*x(1)+k);
q=q+k*cos((k+1)*x(2)+k);
end
fposition=p*q+(x(1)+1.42513)^2+(x(2)+.80032)^2;

The problem of optimization can be solved by two

methods:
Bacteria foraging optimization algorithm
Genetic Algorithm
Genetic algorithm follows the nature‟s law of evolution
Fig. 5. Fitness function plot by Genetic Algorithm
and survival of the fittest. It is based on the behavior of the
chromosomes. The fitter chromosomes survive and the unfit C. Comment
ones don‟t. The decision of the chromosomes survival in the We can see form the above results that the average fitness
next generation is based on the fitness function. This fitness value for a genetic algorithm is -24.279. According to the
function resembles the distance between the bacteria‟s food fitness function selected, lower the value better the
and bacteria. The ones having a lower fitness value for the generation. The best value of the fitness function achieved is
function chosen above will survive in the next generation. -147 approximately.
They are hence capable of mating and producing offspring‟s We can see from the above graphs of BFO that
that may or may not contribute to move towards the goal. As convergence to the minimum point occurs in about four
compared to that in Bacteria foraging optimization generations while in case of GA after execution of around
algorithm, we observe the behavior of the bacteria tumbling 100 generations we get minimum of a function .A
and swimming is expressed as chemotaxis equation which
S.No. Factor Bacterial Genetic
defines the movement of the bacteria. There is a swarming
Foraging Algorithm
characteristic of the bacteria which is taken into account,
Algorithm
where in the bacteria come together in large numbers in
semisolid nutrient medium. 1. Accuracy More Less
On the basis of the above explanations, we observe that 2. Fitness Function -186.565 -147.103
the Bacteria foraging algorithm differs from genetic Value
algorithm. 3. Time 1.2340 sec Around 2
Simulation results using both methods are presented in sec
Fig. 4 and Fig. 5. 4. Optimum point -1.2871 -4.8475
-0.7281 -0.8037
4 INTERNATIONAL CONFERENCE ON CURRENT TRENDS IN TECHNOLOGY, „NUiCONE – 2011‟

comparative analysis of both algorithm is presented in the

following table:

TABLE I
COMPARATIVE ANALYSIS OF BFOA AND GA

D. Conclusion
From above comparative data one can easily understand
that Bacteria foraging optimization algorithm technique is
better than the Genetic algorithm.
Thus Bacterial Foraging algorithm, explains Social
foraging, Genetic Algorithm, Swarm optimization which
thus makes it imperative to analyses strategies required for
Global Optimization.
Optimal Foraging theory uses computational or analytical
methods to provide an optimal foraging policy that specifies
how foraging decisions are made.
Hence the potential uses of Biomimcry of Bacterial
Foraging optimization techniques are to develop adaptive
controllers & co-operative control strategies for autonomous
vehicles.

V. REFERENCES
[1] G. Kevin Passino,”Biomimicry of Bacterial Foraging for Distributed
Optimization and Control,” IEEE Control Systems Magazine, June
2002.
[2] Stephens and J. Krebs, Foraging Theory. Princeton, NJ: Princeton univ.
Press, 1986
[3] W. O‟Brien, H. Browman, and B. Evans, “Search strategies of foraging
animals,” Amer. Scientist,” vol. 78, pp. 152-160, Mar. /Apr. 1990.
[4] H. Berg, Random Walks in Biology. Princeton, NJ: Princeton Univ. Press,
1993.
[5] D. DeRosier, “The turn of the screw: The bacterial flagellar motor,” Cell,
vol. 93, pp. 17-20, 1998.
[6] G. Lowe, M. Meister, and H. Berg, “Rapid rotation of flagellar bundles in
swimming bacteria,” Nature, vol. 325, pp. 637-640, Oct. 1987.
[7] T. Audesirk and G. Audesirk. Biology: Life on Earth. Prentice Hall, NJ,
5 editions, 1999. K. Kristinsson and G. Dumont, “System identification
and control using Genetic algorithms,” IEEE Trans. Syst., Man,
Cybernet., vol. 22, no. 5, pp. 1033-1046, 1992.
[8] Swagatam Das, Sambarta Dasgupta, Arijit Biswas, Ajith Abraham,”On
Stability of the Chemotactic Dynamics in Bacterial-Foraging
Optimization Algorithm, ‟‟ IEEE Trans. Syst.,Man, Cybernet-Part A:
systems and humans, vol. 39, no. 3, may 2009
[9] http://eewww.eng.ohio-state.edu/˜passino