From Wikipedia, the free encyclopedia
In mathematical optimization , the Griewank function is used as a performance test problem for optimization algorithms . For order
n
{\displaystyle n}
it is defined by:[ 1]
f
(
x
1
,
x
2
,
…
,
x
n
)
:=
1
+
1
4000
∑
i
=
1
n
x
i
2
−
∏
i
=
1
n
cos
(
x
i
i
)
.
{\displaystyle f(x_{1},x_{2},\ldots ,x_{n}):=1+{\frac {1}{4000}}\sum _{i=1}^{n}x_{i}^{2}-\prod _{i=1}^{n}\cos \left({\frac {x_{i}}{\sqrt {i}}}\right).}
The Griewank function has a global minimum at
f
(
0
,
0
,
…
,
0
)
=
0
{\displaystyle f(0,0,\ldots ,0)=0}
. A typical search area is
x
i
=
[
−
600
,
600
]
{\displaystyle x_{i}=[-600,600]}
, for
i
=
1
,
…
,
n
{\displaystyle i=1,\ldots ,n}
[ 1] .
First-order Griewank function
g
:=
1
+
(
1
/
4000
)
⋅
x
[
1
]
2
−
cos
(
x
[
1
]
)
{\displaystyle g:=1+(1/4000)\cdot x[1]^{2}-\cos(x[1])}
First order Griewank function has multiple maxima and minima.[ 2]
Let the derivative of Griewank function be zero:
1
2000
⋅
x
[
1
]
+
sin
(
x
[
1
]
)
=
0
{\displaystyle {\frac {1}{2000}}\cdot x[1]+\sin(x[1])=0}
Second-order Griewank function
1
+
1
4000
x
1
2
+
1
4000
x
2
2
−
cos
(
x
1
)
cos
(
1
2
x
2
2
)
{\displaystyle 1+{\frac {1}{4000}}x_{1}^{2}+{\frac {1}{4000}}x_{2}^{2}-\cos(x_{1})\cos \left({\frac {1}{2}}x_{2}{\sqrt {2}}\right)}
Third order Griewank function
{
1
+
1
4000
x
1
2
+
1
4000
x
2
2
+
1
4000
x
3
2
−
cos
(
x
1
)
cos
(
1
2
x
2
2
)
cos
(
1
3
x
3
3
)
}
{\displaystyle \left\{1+{\frac {1}{4000}}\,x_{1}^{2}+{\frac {1}{4000}}\,x_{2}^{2}+{\frac {1}{4000}}\,{x_{3}}^{2}-\cos(x_{1})\cos \left({\frac {1}{2}}x_{2}{\sqrt {2}}\right)\cos \left({\frac {1}{3}}x_{3}{\sqrt {3}}\right)\right\}}
See also
References
^ a b Griewank, A. O. "Generalized Decent for Global Optimization." J. Opt. Th. Appl. 34, 11–39, 1981
^ Locatelli, M. "A Note on the Griewank Test Function." J. Global Opt. 25, 169–174, 2003