es.udc.gii.common.eaf.log.parallel
Class DistributedEvaluationLogTool

java.lang.Object
  extended by es.udc.gii.common.eaf.log.LogTool
      extended by es.udc.gii.common.eaf.log.parallel.ParallelLogTool
          extended by es.udc.gii.common.eaf.log.parallel.DistributedEvaluationLogTool
All Implemented Interfaces:
Configurable, java.util.Observer

public class DistributedEvaluationLogTool
extends ParallelLogTool

Logs the evaluations in a parallel evolutionary algorithm. The output for a master node reads as follows (although the sequence might differ because of parallel processing):

 < time ms > [topology rank ... rank of the node ... ]
 TOTAL INDIVIDUALS TO EVALUATE:
 ... list of individuals to evaluate in this generation ...
 
 < time ms > [topology rank ... rank of the node ... ]
 Sending EvaluationObject:
 ... evaluation object sent to slave or to self ...

 < time ms > [topology rank ... rank of the node ... ]
 Received EvaluationObject:
 ... evaluation object received from slave or from self ...

 < time ms > [topology rank ... rank of the node ... ]
 Individuals to evaluate:
 ... chunk of individuals to evaluate ...

 < time ms > [topology rank ... rank of the node ... ]
 Evaluated individuals:
 ... chunk of evaluated individuals ...

 < time ms > [topology rank ... rank of the node ... ]
 STATS:
 ... statistics collected by the master node ...

 < time ms > [topology rank ... rank of the node ... ]
 TOTAL EVALUATED INDIVIDUALS:
 ... total evaluated individuals in this generation ...

 -------------------
 
Several send and receive are logged as the distribution of the evaluation goes on. The statistics are only print if the distributed evaluation strategy has been configured for collecting them (see DistributedEvaluation).

The output for a slave node reads as follows (although the sequence might differ because of parallel processing):

 < time ms > [topology rank ... rank of the node ... ]
 Received EvaluationObject:
 ... evaluation object received from slave or from self ...

 < time ms > [topology rank ... rank of the node ... ]
 Individuals to evaluate:
 ... chunk of individuals to evaluate ...

 < time ms > [topology rank ... rank of the node ... ]
 Evaluated individuals:
 ... chunk of evaluated individuals ...
 
 < time ms > [topology rank ... rank of the node ... ]
 Sending EvaluationObject:
 ... evaluation object sent to slave or to self ...

 < time ms > [topology rank ... rank of the node ... ]
 STATS:
 ... statistics collected by the master node ...
 
Several send and receive are logged as the distribution of the evaluation goes on. The statistics are only print if the distributed evaluation strategy has been configured for collecting them (see DistributedEvaluation).

Since:
1.0
Author:
Grupo Integrado de IngenierĂ­a (www.gii.udc.es)
See Also:
DistributedEvaluation

Field Summary
 
Fields inherited from class es.udc.gii.common.eaf.log.parallel.ParallelLogTool
t0
 
Fields inherited from class es.udc.gii.common.eaf.log.LogTool
fileExtension, name, oldFolder, oldName
 
Constructor Summary
DistributedEvaluationLogTool()
          Creates a new instance of DistributedEvaluationLogTool
 
Method Summary
 java.lang.String getLogID()
           
 void update(java.util.Observable o, java.lang.Object arg)
           
 
Methods inherited from class es.udc.gii.common.eaf.log.parallel.ParallelLogTool
getNodeID
 
Methods inherited from class es.udc.gii.common.eaf.log.LogTool
configure, finalize, getLog, setFile
 
Methods inherited from class java.lang.Object
clone, equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Constructor Detail

DistributedEvaluationLogTool

public DistributedEvaluationLogTool()
Creates a new instance of DistributedEvaluationLogTool

Method Detail

getLogID

public java.lang.String getLogID()
Overrides:
getLogID in class LogTool

update

public void update(java.util.Observable o,
                   java.lang.Object arg)
Specified by:
update in interface java.util.Observer
Overrides:
update in class LogTool