Genetic Algorithms

Special assignment 3 for AI course at ITBA

Building

This framework uses maven. To build it:

$ git clone https://github.com/lmarzora/SIA-TP3.git master
$ cd SIA-TP3/
$ mvn clean package
$ java -jar target/sia-tp3-1.0-SNAPSHOT-jar-with-dependencies.jar [-config]

Config File

This repo comes with a default configuration file default.properties. You can supply your own by using the -config key when running the program.

Global

Parameter	Value in config file	Example
Gear Data	global.data	fulldata
A	global.a	0.3
B	global.b	0.3
N	global.N	1000

Character Features

Parameter	Value in config file	Example
Heir	multi.heir	warrior
Strength	multi.strength	0.8
Dexterity	multi.dexterity	0.9
Expertise	multi.expertise	0.9
Resistance	multi.resistance	1.2
Life	multi.life	1.1

Cut Conditions

Parameter	Value in config file	Example
Max number of generations	cut.maxgenerations	10000
Percentage of unchanged population between two generations	cut.structure	0.8
Number of generations to evaluate if fitness doesn't improve (using mean fitness)	cut.contentmean	1500
Number of generations to evaluate if fitness doesn't improve (using max fitness)	cut.contentmax	3000
Maximum fitness wanted	cut.maxfitness	30

NOTE: If this parameter is present, the cut condition will be enforced. For example, if the config file has the cut.contentmax and cut.structure keys, the cut conditions applied are Content Max and Structure.

Selection

Parameter	Value in config file	Example
Selection Temperature	selection.temperature	0.25
Selection Children Size	selection.m	2
Selection Probability	selection.probability	0.75
Parent Selection Method 1	selection.parentMethod1	elite
Parent Selection Method 2	selection.parentMethod2	boltzmann
Children Selection Method 1	selection.childrenMethod1	elite
Children Selection Method 2	selection.childrenMethod2	roulette

Possible method values

Method	Key
Boltzmann Selector	boltzmann
Deterministic Tournament	deterministic
Elite Selector	elite
Random Distinct Selector	randomd
Random Selector	random
Ranking Selector	ranking
Roulette Selector	roulette
Stochastic Tournament	stochastic
Universal Selector	universal

Replacement

Parameter	Value in config file	Example
Replacement Children Size	replacement.k	500
Replacement Method	replacement.method	batch

Possible method values

Method	Key
Iterative Replacer (Method 1)	iterative
Batch Replacer (Method 2)	batch
Generational Replacer (Method 3)	generational

Reproduction

Parameter	Value in config file	Example
Reproduction Method	reproduction.method	anular

Possible method values

Method	Key
Anular Cross	anular
One Point Cross	onepoint
Two Point Cross	twopoints
Uniform Cross	uniform

Mutation

Parameter	Value in config file	Example
Mutation Probability	mutation.probability	0.3
Mutation Minimum Height	mutation.minHeight	1.3
Mutation Maximum Height	mutation.maxHeight	2.0
Mutation Method	mutation.method	uniform

Possible method values

Method	Key
Uniform Mutator	uniform

Credits

• Marzoratti, Luis
• Soncini, Lucas
• Fraga, Matias
• De Lucca, Tomas