<s>
In	O
computational	B-General_Concept
intelligence	I-General_Concept
(	O
CI	O
)	O
,	O
an	O
evolutionary	B-Algorithm
algorithm	I-Algorithm
(	O
EA	O
)	O
is	O
a	O
subset	O
of	O
evolutionary	O
computation	O
,	O
a	O
generic	O
population-based	O
metaheuristic	B-Algorithm
optimization	O
algorithm	O
.	O
</s>
<s>
An	O
EA	O
uses	O
mechanisms	O
inspired	O
by	O
biological	O
evolution	O
,	O
such	O
as	O
reproduction	O
,	O
mutation	O
,	O
recombination	O
,	O
and	O
selection	B-Application
.	O
</s>
<s>
Evolutionary	B-Algorithm
algorithms	I-Algorithm
often	O
perform	O
well	O
approximating	O
solutions	O
to	O
all	O
types	O
of	O
problems	O
because	O
they	O
ideally	O
do	O
not	O
make	O
any	O
assumption	O
about	O
the	O
underlying	O
fitness	O
landscape	O
.	O
</s>
<s>
Techniques	O
from	O
evolutionary	B-Algorithm
algorithms	I-Algorithm
applied	O
to	O
the	O
modeling	O
of	O
biological	O
evolution	O
are	O
generally	O
limited	O
to	O
explorations	O
of	O
microevolutionary	O
processes	O
and	O
planning	O
models	O
based	O
upon	O
cellular	O
processes	O
.	O
</s>
<s>
Fitness	B-Algorithm
approximation	I-Algorithm
is	O
one	O
of	O
the	O
solutions	O
to	O
overcome	O
this	O
difficulty	O
.	O
</s>
<s>
The	O
following	O
is	O
an	O
example	O
of	O
a	O
generic	O
single-objective	O
genetic	B-Algorithm
algorithm	I-Algorithm
.	O
</s>
<s>
Step	O
Two	O
:	O
Repeat	O
the	O
following	O
regenerational	O
steps	O
until	O
termination	B-Application
:	O
</s>
<s>
Similar	O
techniques	O
differ	O
in	O
genetic	B-Algorithm
representation	I-Algorithm
and	O
other	O
implementation	O
details	O
,	O
and	O
the	O
nature	O
of	O
the	O
particular	O
applied	O
problem	O
.	O
</s>
<s>
Genetic	B-Algorithm
algorithm	I-Algorithm
–	O
This	O
is	O
the	O
most	O
popular	O
type	O
of	O
EA	O
.	O
</s>
<s>
Genetic	B-Algorithm
programming	I-Algorithm
–	O
Here	O
the	O
solutions	O
are	O
in	O
the	O
form	O
of	O
computer	O
programs	O
,	O
and	O
their	O
fitness	O
is	O
determined	O
by	O
their	O
ability	O
to	O
solve	O
a	O
computational	O
problem	O
.	O
</s>
<s>
There	O
are	O
many	O
variants	O
of	O
Genetic	B-Algorithm
Programming	I-Algorithm
,	O
including	O
Cartesian	B-Algorithm
genetic	I-Algorithm
programming	I-Algorithm
,	O
gene	B-Algorithm
expression	I-Algorithm
programming	I-Algorithm
,	O
grammatical	B-Algorithm
evolution	I-Algorithm
,	O
linear	B-Algorithm
genetic	I-Algorithm
programming	I-Algorithm
,	O
multi	B-Algorithm
expression	I-Algorithm
programming	I-Algorithm
etc	O
.	O
</s>
<s>
Evolutionary	B-Algorithm
programming	I-Algorithm
–	O
Similar	O
to	O
genetic	B-Algorithm
programming	I-Algorithm
,	O
but	O
the	O
structure	O
of	O
the	O
program	O
is	O
fixed	O
and	O
its	O
numerical	O
parameters	O
are	O
allowed	O
to	O
evolve	O
.	O
</s>
<s>
Evolution	B-Algorithm
strategy	I-Algorithm
–	O
Works	O
with	O
vectors	O
of	O
real	O
numbers	O
as	O
representations	O
of	O
solutions	O
,	O
and	O
typically	O
uses	O
self-adaptive	O
mutation	O
rates	O
.	O
</s>
<s>
Differential	B-Algorithm
evolution	I-Algorithm
–	O
Based	O
on	O
vector	O
differences	O
and	O
is	O
therefore	O
primarily	O
suited	O
for	O
numerical	O
optimization	O
problems	O
.	O
</s>
<s>
Neuroevolution	B-Algorithm
–	O
Similar	O
to	O
genetic	B-Algorithm
programming	I-Algorithm
but	O
the	O
genomes	O
represent	O
artificial	O
neural	O
networks	O
by	O
describing	O
structure	O
and	O
connection	O
weights	O
.	O
</s>
<s>
The	O
genome	O
encoding	B-Algorithm
can	O
be	O
direct	O
or	O
indirect	O
.	O
</s>
<s>
Learning	B-Algorithm
classifier	I-Algorithm
system	I-Algorithm
–	O
Here	O
the	O
solution	O
is	O
a	O
set	O
of	O
classifiers	O
(	O
rules	O
or	O
conditions	O
)	O
.	O
</s>
<s>
Initially	O
,	O
classifiers	O
were	O
only	O
binary	O
,	O
but	O
now	O
include	O
real	O
,	O
neural	O
net	O
,	O
or	O
S-expression	B-Protocol
types	O
.	O
</s>
<s>
Fitness	O
is	O
typically	O
determined	O
with	O
either	O
a	O
strength	O
or	O
accuracy	O
based	O
reinforcement	O
learning	O
or	O
supervised	B-General_Concept
learning	I-General_Concept
approach	O
.	O
</s>
<s>
Under	O
the	O
same	O
condition	O
,	O
no	O
evolutionary	B-Algorithm
algorithm	I-Algorithm
is	O
fundamentally	O
better	O
than	O
another	O
.	O
</s>
<s>
by	O
choosing	O
a	O
certain	O
mutation	O
strength	O
or	O
a	O
problem-adapted	B-Algorithm
coding	I-Algorithm
)	O
.	O
</s>
<s>
In	O
addition	O
,	O
an	O
EA	O
can	O
use	O
problem	O
specific	O
knowledge	O
by	O
,	O
for	O
example	O
,	O
not	O
randomly	O
generating	O
the	O
entire	O
start	O
population	O
,	O
but	O
creating	O
some	O
individuals	O
through	O
heuristics	B-Algorithm
or	O
other	O
procedures	O
.	O
</s>
<s>
Another	O
possibility	O
to	O
tailor	O
an	O
EA	O
to	O
a	O
given	O
problem	O
domain	O
is	O
to	O
involve	O
suitable	O
heuristics	B-Algorithm
,	O
local	B-Algorithm
search	I-Algorithm
procedures	I-Algorithm
or	O
other	O
problem-related	O
procedures	O
in	O
the	O
process	O
of	O
generating	O
the	O
offspring	O
.	O
</s>
<s>
This	O
form	O
of	O
extension	O
of	O
an	O
EA	O
is	O
also	O
known	O
as	O
a	O
memetic	B-Algorithm
algorithm	I-Algorithm
.	O
</s>
<s>
For	O
EAs	O
in	O
which	O
,	O
in	O
addition	O
to	O
the	O
offspring	O
,	O
at	O
least	O
the	O
best	O
individual	O
of	O
the	O
parent	O
generation	O
is	O
used	O
to	O
form	O
the	O
subsequent	O
generation	O
(	O
so-called	O
elitist	O
EAs	O
)	O
,	O
there	O
is	O
a	O
general	O
proof	O
of	O
convergence	B-Application
under	O
the	O
condition	O
that	O
an	O
optimum	O
exists	O
.	O
</s>
<s>
I.e.	O
,	O
the	O
fitness	O
values	O
represent	O
a	O
monotonically	O
non-decreasing	O
sequence	O
,	O
which	O
is	O
bounded	B-Algorithm
due	O
to	O
the	O
existence	O
of	O
the	O
optimum	O
.	O
</s>
<s>
From	O
this	O
follows	O
the	O
convergence	B-Application
of	O
the	O
sequence	O
against	O
the	O
optimum	O
.	O
</s>
<s>
Since	O
the	O
proof	O
makes	O
no	O
statement	O
about	O
the	O
speed	O
of	O
convergence	B-Application
,	O
it	O
is	O
of	O
little	O
help	O
in	O
practical	O
applications	O
of	O
EAs	O
.	O
</s>
<s>
In	O
a	O
panmictic	O
population	O
model	O
,	O
mate	O
selection	B-Application
(	O
step	O
2	O
of	O
the	O
section	O
about	O
implementation	O
)	O
is	O
such	O
that	O
every	O
individual	O
in	O
the	O
entire	O
population	O
is	O
eligible	O
as	O
a	O
mate	O
.	O
</s>
<s>
In	O
non-panmictic	O
populations	O
,	O
selection	B-Application
is	O
suitably	O
restricted	O
,	O
so	O
that	O
the	O
dispersal	O
speed	O
of	O
better	O
individuals	O
is	O
reduced	O
compared	O
to	O
panmictic	O
ones	O
.	O
</s>
<s>
Thus	O
,	O
the	O
general	O
risk	O
of	O
premature	O
convergence	B-Application
of	O
elitist	O
EAs	O
can	O
be	O
significantly	O
reduced	O
by	O
suitable	O
population	O
models	O
that	O
restrict	O
mate	O
selection	B-Application
.	O
</s>
<s>
A	O
possible	O
limitation	O
of	O
many	O
evolutionary	B-Algorithm
algorithms	I-Algorithm
is	O
their	O
lack	O
of	O
a	O
clear	O
genotype	O
–	O
phenotype	O
distinction	O
.	O
</s>
<s>
This	O
indirect	O
encoding	B-Algorithm
is	O
believed	O
to	O
make	O
the	O
genetic	O
search	O
more	O
robust	O
(	O
i.e.	O
</s>
<s>
reduce	O
the	O
probability	O
of	O
fatal	O
mutations	O
)	O
,	O
and	O
also	O
may	O
improve	O
the	O
evolvability	B-General_Concept
of	O
the	O
organism	O
.	O
</s>
<s>
Such	O
indirect	O
(	O
also	O
known	O
as	O
generative	O
or	O
developmental	O
)	O
encodings	B-Algorithm
also	O
enable	O
evolution	O
to	O
exploit	O
the	O
regularity	O
in	O
the	O
environment	O
.	O
</s>
<s>
Recent	O
work	O
in	O
the	O
field	O
of	O
artificial	B-Algorithm
embryogeny	I-Algorithm
,	O
or	O
artificial	O
developmental	O
systems	O
,	O
seeks	O
to	O
address	O
these	O
concerns	O
.	O
</s>
<s>
And	O
gene	B-Algorithm
expression	I-Algorithm
programming	I-Algorithm
successfully	O
explores	O
a	O
genotype	O
–	O
phenotype	O
system	O
,	O
where	O
the	O
genotype	O
consists	O
of	O
linear	O
multigenic	O
chromosomes	O
of	O
fixed	O
length	O
and	O
the	O
phenotype	O
consists	O
of	O
multiple	O
expression	O
trees	O
or	O
computer	O
programs	O
of	O
different	O
sizes	O
and	O
shapes	O
.	O
</s>
<s>
The	O
areas	O
in	O
which	O
evolutionary	B-Algorithm
algorithms	I-Algorithm
are	O
practically	O
used	O
are	O
almost	O
unlimited	O
and	O
range	O
from	O
industry	O
,	O
engineering	O
,	O
complex	O
scheduling	O
,	O
agriculture	O
,	O
robot	O
movement	O
planning	O
and	O
finance	O
to	O
research	O
and	O
art	B-Algorithm
.	O
</s>
<s>
The	O
application	O
of	O
an	O
evolutionary	B-Algorithm
algorithm	I-Algorithm
requires	O
some	O
rethinking	O
from	O
the	O
inexperienced	O
user	O
,	O
as	O
the	O
approach	O
to	O
a	O
task	O
using	O
an	O
EA	O
is	O
different	O
from	O
conventional	O
exact	O
methods	O
and	O
this	O
is	O
usually	O
not	O
part	O
of	O
the	O
curriculum	O
of	O
engineers	O
or	O
other	O
disciplines	O
.	O
</s>
<s>
Swarm	B-Architecture
algorithms	I-Architecture
include	O
:	O
</s>
<s>
Ant	B-Algorithm
colony	I-Algorithm
optimization	I-Algorithm
is	O
based	O
on	O
the	O
ideas	O
of	O
ant	O
foraging	O
by	O
pheromone	O
communication	O
to	O
form	O
paths	O
.	O
</s>
<s>
Artificial	B-Algorithm
bee	I-Algorithm
colony	I-Algorithm
algorithm	I-Algorithm
is	O
based	O
on	O
the	O
honeybee	O
foraging	O
behaviour	O
.	O
</s>
<s>
Bees	B-Algorithm
algorithm	I-Algorithm
is	O
based	O
on	O
the	O
foraging	O
behaviour	O
of	O
honeybees	O
.	O
</s>
<s>
Cuckoo	B-Algorithm
search	I-Algorithm
is	O
inspired	O
by	O
the	O
brooding	O
parasitism	O
of	O
the	O
cuckoo	O
species	O
.	O
</s>
<s>
Particle	B-Algorithm
swarm	I-Algorithm
optimization	I-Algorithm
is	O
based	O
on	O
the	O
ideas	O
of	O
animal	O
flocking	O
behaviour	O
.	O
</s>
<s>
Adaptive	B-Algorithm
dimensional	I-Algorithm
search	I-Algorithm
–	O
Unlike	O
nature-inspired	O
metaheuristic	B-Algorithm
techniques	O
,	O
an	O
adaptive	B-Algorithm
dimensional	I-Algorithm
search	I-Algorithm
algorithm	O
does	O
not	O
implement	O
any	O
metaphor	O
as	O
an	O
underlying	O
principle	O
.	O
</s>
<s>
Firefly	B-Algorithm
algorithm	I-Algorithm
is	O
inspired	O
by	O
the	O
behavior	O
of	O
fireflies	O
,	O
attracting	O
each	O
other	O
by	O
flashing	O
light	O
.	O
</s>
<s>
Memetic	B-Algorithm
algorithm	I-Algorithm
–	O
A	O
hybrid	O
method	O
,	O
inspired	O
by	O
Richard	O
Dawkins	O
's	O
notion	O
of	O
a	O
meme	O
,	O
it	O
commonly	O
takes	O
the	O
form	O
of	O
a	O
population-based	O
algorithm	O
coupled	O
with	O
individual	O
learning	O
procedures	O
capable	O
of	O
performing	O
local	O
refinements	O
.	O
</s>
<s>
In	O
2020	O
,	O
Google	B-Application
stated	O
that	O
their	O
AutoML-Zero	O
can	O
successfully	O
rediscover	O
classic	O
algorithms	O
such	O
as	O
the	O
concept	O
of	O
neural	O
networks	O
.	O
</s>
