The MESA Program: An Executive
Summary
By Micah Sparacio
The
(very)basic algorithm:
1. Randomly generate initial
population
2. Apply fitness function
in which couplings are only given a beneficial fitness value if
each bit in the coupling matches the target. Then apply the fitness
perturbation if it is being used.
3. Re-populate: if elitism
is on, keep the most fit half of population the same, and reproduce
new half with mutation from the most fit half. If elitism is off,
take most fit half of population, and reproduce each one twice with
mutation.
4. If crossover is on, do
crossovers based on crossover rate and randomly generated crossover
pairs
5. Do steps 2-4 until you
have reached the target organism
Some
guidelines for use:
Please note that at this
stage error checking on user input is not yet implemented. It is
up to you the user to input the appropriate values.
This program allows for the
user to modify the following parameters:
1. Mutation Distribution
Type
- Bit by bit: probability
assessed at each bit in genome (local)
- Poisson Distribution: probability over entire population (global)
- at Peak: moving toward
higher fitness value
- at Valley: moving toward lower fitness value
-Uniform Distribution:
random number between -X and +X
-Binomial Distribution: a binomial distribution using n and p values
-Manual Distribution: perturbation determined by values input at
bottom of screen
-On: Most fit organisms
persist without mutation
-Off: Most fit organisms are subject to mutation
-On: Organisms subject
to crossover rate
-Off: Organisms not subject to crossover rate
6. Population size:
number of organisms
7. String Length:
bit length of organisms
8. Uniform Fitness
Pert. Range: used if Fitness Perturbation type is set to
Uniform Distribution
9. Binomial Fitness
Pert.: used if Fitness Perturbation type is set to Binary
Distribution: if n=20 and p=.5 then you will get the equivalent
of 20 random coin tosses with each toss having a 50/50 probability
of landing heads. If you increase the p value, the probability of
landing heads goes up.
10. Mutation Rate:
one mutation for every X bits. Used by both types of Mutation Distribution
11. Crossover rate:
one crossover for every X bit pairs
12. Coupling Degree/Total
Couplings Pairings
- Coupling degree is the
amount of bits that are grouped together
- Total Couplings is the number of blocks in each organism with
the specified coupling degree
13. Deviation/Distribution
Pairings
-Deviation is a positive
or negative integer value perturbation of the fitness value
- % Distribution is the percentage of time the deviation value will
be used. Deviation values must add up to 100 percent
14. Keep Record Check
Box and Text Box
-Specify whether to keep
a record of crossovers (crossovers.txt), coupled bits (coupledbits.txt),
and new best fitness values (bestfitness.txt). The files will be
stored in the same directory that the program is run from
-Specify the number of generations between records. The higher the
number, the fewer records kept, the less computational time required,
and the smaller the files stored.
|
