Information Technology Reference
In-Depth Information
It has a fitness of 10.7883 and an R-square of 0.9259986046 evaluated over
the training set of 90 fitness cases and, thus, is slightly better than the model
(7.9) created without random numerical constants and considerably better
than all the models designed with Kolmogorov-Gabor polynomials. More
formally, the model (7.10a) can be expressed by the following function:
0
504425
c
(
g
i
)
(
a
c
)
(
j
i
)
y
0
849945
j
(7.10b)
3
i
a
2
i
b
In the multicellular system with ADFs without random numerical con-
stants, the best solution was found in generation 1996 of run 2 (the neutral
cells are not shown):
012345678901234
i*f++//ajghdbje
+/jb-bfdbhehfei
+*j/a++hcjibhcc
/*++1+022021010
(7.11a)
It has a fitness of 8.63869 and an R-square of 0.9100736178 evaluated over
the training set of 90 fitness cases and, therefore, is also considerably better
than all the models designed with the STROGANOFF systems. More for-
mally, the model (7.11a) above can be expressed by the following function
(the contribution of each ADF is shown in brackets):
§
h
c
·
§
b
·
¨
©
¸
¹
¨
©
¸
¹
2
a
j
j
j
i
b
f
y
(7.11b)
§
h
c
·
¨
©
¸
¹
2
(
i
)
a
j
j
i
In the multicellular system with RNCs, the best solution was designed in
run 71 after 4561 generations (the neutral cells are not shown):
Gene 0: +j/++-+egajhhj?21492592
C
0
:
{0.631439, -0.25534, -0.833344, 0.831757, -0.645355,
-0.542816, -0.72107, 0.228332, -0.170013, -0.259155}
Gene 1: i/bd/-*ghfgjjde86346814
C
1
:
{0.369568, 0.184173, -0.173675, 0.604859, -0.843994,
-0.537872, -0.44339, 0.228332, -0.951661, -0.916932}
Gene 2: *+?i**??didg?aj77428535
C
2
:
{-0.306518, -0.864869, -0.668701, -0.181671, 0.701142,
0.77948, -0.183654, -0.518463, 0.567505, -0.48941}
Homeotic gene: /*+0+-100022110
(7.12a)
