Bio-inspired Evolutionary Computing with Context-Awareness and Collective-Effect - Technologies and Applications of Artificial Intelligence - page 108

4 Simulation Resu ults

Two benchmark functions a
rithm [16]. These selected b

are used to validate the performance of the proposed al
benchmark functions are described in details as follows.

lgo-

• Griewangk Function

This function is similar to
minima distributed regularl

the Rastrigin function. There are many widespread lo
y as shown in Fig. 2. It is defined as follows,

ocal

1 cos

4000

1

√

600 600, 1,…,.

(2 21)

The global minimum
191 local minimal as d=1 i
increases with d . When the
solution becomes extremel
minima is more than the Ra
whole solution space. The
proposed algorithm will be

0 is obtainable when 0 , i=1,…,n . There
n this function. The number of local minima exponentia
number of local minima fast increases, the global optim
ly difficult to detect as d increment. The number of lo
astrigin function. And the local minimal are distributive
ability of escapement from local optimal solution for
demonstrated by this function.

are
ally
mal
ocal
e in
the

Fig. . 2. Griewangk function in 2-dimension

• Ackley Function

Ackley function is based o
modal as shown in Fig. 3. T
modulated with an exponen
outer region and a control
come more and more influe

on the gradient and widely used as test function of mu
This function is with a cosine wave of moderate amplit
ntial function. Ackley function is composed of a plane-l
hole or peak where modulations by the cosine wave
ential. A typical Ackley function is as follows,

ulti-
tude
like
be-