|
|||||||||
PREV CLASS NEXT CLASS | FRAMES NO FRAMES | ||||||||
SUMMARY: INNER | FIELD | CONSTR | METHOD | DETAIL: FIELD | CONSTR | METHOD |
java.lang.Object | +--ec.BreedingSource | +--ec.SelectionMethod | +--ec.parsimony.RatioBucketTournament
Does a tournament selection, limited to the subpopulation it's working in at the time.
Bucket Lexicographic Tournament selection works like as follows. The sizes of buckets are proportioned so that low-fitness individuals are placed into much larger buckets than high-fitness individuals. A bucket ratio 1/ratio is specified beforehand. The bottom 1/ratio individuals of the population are placed into the bottom bucket. If any individuals remain in the population with the same fitness as the best individual in the bottom bucket, they too are placed in that bucket. Of the remaining population, the next 1/ratio individuals are placed into the next bucket, plus any individuals remaining in the population with the same fitness as the best individual now in that bucket, and so on. This continues until every member of the population has been placed in a bucket. Once again, the fitness of every individual in a bucket is set to the rank of the bucket relative to other buckets. Ratio bucketing thus allows parsimony to have more of an effect on average when two similar low-fitness individuals are considered than when two high-fitness individuals are considered. After ranking the individuals, size individuals are chosen at random from the population. Of those individuals, the one with the highest rank is selected. If the two individuals are in the same rank, meaning that they have similar fitness, the one with the smallest size is selected.
Bucket Lexicographic Tournament selection is so simple that it doesn't need to maintain a cache of any form, so many of the SelectionMethod methods just don't do anything at all.
Typical Number of Individuals Produced Per produce(...) call
Always 1.
Parameters
base.size int >= 1 (default 7) |
(the tournament size) |
base.pick-worst bool = true or false (default) |
(should we pick the worst individual in the tournament instead of the best?) |
base.ratio float >= 2 (default) |
(the ratio of worst out of remaining individuals that go in the next bucket) |
Default Base
select.ratio-bucket-tournament
Field Summary | |
static int |
DEFAULT_SIZE
Default size |
static java.lang.String |
P_PICKWORST
if the worst individual should be picked in the tournament |
static java.lang.String |
P_RATIO
The value of RATIO: each step, the worse 1/RATIO individuals are assigned the same fitness |
static java.lang.String |
P_SIZE
size parameter |
static java.lang.String |
P_TOURNAMENT
default base |
boolean |
pickWorst
Do we pick the worst instead of the best? |
float |
ratio
The value of RATIO |
int |
size
Size of the tournament |
Fields inherited from class ec.SelectionMethod |
INDS_PRODUCED |
Fields inherited from class ec.BreedingSource |
CHECKBOUNDARY, DEFAULT_PRODUCED, NO_PROBABILITY, P_PROB, probability, UNUSED |
Constructor Summary | |
RatioBucketTournament()
|
Method Summary | |
Parameter |
defaultBase()
Returns the default base for this prototype. |
void |
individualReplaced(SteadyStateEvolutionState state,
int subpopulation,
int thread,
int individual)
Called whenever an individual has been replaced by another in the population. |
void |
prepareToProduce(EvolutionState state,
int subpopulation,
int thread)
Prepare to produce: create the buckets!!!! |
int |
produce(int subpopulation,
EvolutionState state,
int thread)
An alternative form of "produce" special to Selection Methods; selects an individual from the given subpopulation and returns its position in that subpopulation. |
void |
setup(EvolutionState state,
Parameter base)
Sets up the BreedingPipeline. |
void |
sourcesAreProperForm(SteadyStateEvolutionState state)
Issue an error (not a fatal -- we guarantee that callers of this method will also call exitIfErrors) if any of your sources, or their sources, etc., are not of SteadyStateBSourceForm. |
Methods inherited from class ec.SelectionMethod |
finishProducing, preparePipeline, produce, produces, typicalIndsProduced |
Methods inherited from class ec.BreedingSource |
getProbability, pickRandom, protoClone, protoCloneSimple, setProbability, setupProbabilities |
Methods inherited from class java.lang.Object |
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait |
Field Detail |
public static final java.lang.String P_TOURNAMENT
public static final java.lang.String P_SIZE
public static final int DEFAULT_SIZE
public int size
public static final java.lang.String P_PICKWORST
public boolean pickWorst
public static final java.lang.String P_RATIO
public float ratio
Constructor Detail |
public RatioBucketTournament()
Method Detail |
public Parameter defaultBase()
Prototype
public void setup(EvolutionState state, Parameter base)
BreedingSource
The most common modification is to normalize it with some other set of probabilities, then set all of them up in increasing summation; this allows the use of the fast static BreedingSource-picking utility method, BreedingSource.pickRandom(...). In order to use this method, for example, if four breeding source probabilities are {0.3, 0.2, 0.1, 0.4}, then they should get normalized and summed by the outside owners as: {0.3, 0.5, 0.6, 1.0}.
setup
in class BreedingSource
ec.BreedingSource
Prototype.setup(EvolutionState,Parameter)
public void prepareToProduce(EvolutionState state, int subpopulation, int thread)
prepareToProduce
in class SelectionMethod
public int produce(int subpopulation, EvolutionState state, int thread)
SelectionMethod
produce
in class SelectionMethod
public void individualReplaced(SteadyStateEvolutionState state, int subpopulation, int thread, int individual)
SteadyStateBSourceForm
individualReplaced
in interface SteadyStateBSourceForm
public void sourcesAreProperForm(SteadyStateEvolutionState state)
SteadyStateBSourceForm
sourcesAreProperForm
in interface SteadyStateBSourceForm
|
|||||||||
PREV CLASS NEXT CLASS | FRAMES NO FRAMES | ||||||||
SUMMARY: INNER | FIELD | CONSTR | METHOD | DETAIL: FIELD | CONSTR | METHOD |