Checkpoint 6 -- Population Whatyouwillbedoing Goals: - PowerPoint PPT Presentation

Checkpoint 6 -- Population

What
you
will
be
doing
 • Goals:
 • Experiment
with
different
crossover
/
mutation
 rates
 • Experiment
with
a
variety
of
evolutionary
strategies
 for
controlling
population
 • More
statistics
gathering.
 • Reminder:
all
checkpoints
to
contribute
to
what
 eventually
will
be
reported
in
your
final
report
/
 presentation.


Deliverables
 • Report
and
Code


Reproduction
 Generation
k+1
 Generation
k
 Reproduction


Selection
 • Crossover
 • Compare:
 • Crossover
rate:
0
 • Crossover
rate
=
0.25
 • Crossover
rate
=
0.50
 • Crossover
rate
=
0.75
 • Crossover
rate
=
1.0
 • Maintain
0.001
mutation
rate


Selection
 • Mutation
 • Compare:
 • Mutation
rate:
0.001
 • Mutation
rate
0.01
 • Mutation
rate
0.05
 • Maintain
75%
crossover
rate.


Population
Parameters
 • Tasks:
 • Choose
2
of
the
3
evolutionary
strategies
 • ES
( µ 
+
 λ )
 • Steady
State

 • Variable
Population
 • Or
 • Choose
3
population
sizes
 • Small,
medium,
and
large
 • Compare
EA
convergence
of
using
each.


Comparisons
 • Keep
all
other
parameters
constant
 • Crossover
/
mutation
operators
 • Selection
Mechanism
 • Crossover
/
mutation
rate.
 • Genotype
/
Phenotype
/
Genetic
Mapping
 • Use
values
that
showed
most
promise
in
 previous
checkpoints.


Report
 • Be
sure
to
include:
 • Which
evolutionary
strategies
you
chose
(with
 descriptions)
 • Include:
 • Summary
of
runs
(best,
worst,
avg
/
generation)
 • Conclusions
on
which
seems
most
promising.


Questions?


Ground
rules
 • Report
submission
in
PDF,
Word,
or
plain
text.
 • Code
submission
as
zip,
tar,
etc.
 • Include
instructions
for
building/running.
 • Include
platform
as
mycourses
comment
when
 submitting.
 • Electronic
submission
via
mycourses.


Submission
 • Due
Tuesday,
November
3 rd 

 • Any
trouble,
see
me
sooner
rather
than
later.


Reminder
 • This
is
our
final
checkpoint
 • All
findings
will
contribute
to
final
report.
 • More
on
final
report
/
presentation
next
week