Package org.pentaho.aggdes.algorithm.impl

Class AdaptiveMonteCarloAlgorithm

java.lang.Object

org.pentaho.aggdes.algorithm.impl.AlgorithmImpl

org.pentaho.aggdes.algorithm.impl.AdaptiveMonteCarloAlgorithm

All Implemented Interfaces:

Algorithm, Component

public class AdaptiveMonteCarloAlgorithm
extends AlgorithmImpl

Implementation of Algorithm that runs the Monte Carlo method repeatedly, tuning the cost limit each time to yield an effective cost limit.

The Monte Carlo algorithm tends to be most effective when the target cost is close to the actual cost. By running the algorithm repeatedly, we can determine a realistic cost.

Nested Class Summary

Nested classes/interfaces inherited from interface org.pentaho.aggdes.algorithm.Algorithm

Algorithm.CostBenefit, Algorithm.ParameterEnum

Field Summary

Fields inherited from class org.pentaho.aggdes.algorithm.impl.AlgorithmImpl

parameterList, schema

Constructor Summary

Constructors

Constructor	Description
AdaptiveMonteCarloAlgorithm()

Method Summary

Modifier and Type	Method	Description
List<Algorithm.CostBenefit>	computeAggregateCosts(Schema schema, Map<Parameter,Object> parameterValues, List<Aggregate> aggregateList)	Computes cost/benefit metrics for a list of Aggregates.
Result	run(Schema schema, Map<Parameter,Object> parameterValues, Progress progress)	Runs the optimization algorithm to produces a set of recommended aggregate tables.

Methods inherited from class org.pentaho.aggdes.algorithm.impl.AlgorithmImpl

cancel, checkCancelTimeout, createAggregate, getBaseName, getName, getParameters, onStart, runAlgorithm

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

AdaptiveMonteCarloAlgorithm

public AdaptiveMonteCarloAlgorithm()

Method Details

run

public Result run(Schema schema,
 Map<Parameter,Object> parameterValues,
 Progress progress)

Description copied from interface: Algorithm

Runs the optimization algorithm to produces a set of recommended aggregate tables.

The allowable parameters are specified by the Component.getParameters() method.

Returns null if another thread called Algorithm.cancel() and there is no useful result.

Parameters:

schema - Schema

parameterValues - Parameter values

progress - Progress callback

Returns:

Set of recommended aggregate tables, or null if another thread called Algorithm.cancel() and there is no useful result

computeAggregateCosts

public List<Algorithm.CostBenefit> computeAggregateCosts(Schema schema,
 Map<Parameter,Object> parameterValues,
 List<Aggregate> aggregateList)

Description copied from interface: Algorithm

Computes cost/benefit metrics for a list of Aggregates.

The aggregates must have been created using the Algorithm.createAggregate(Schema, List) method.

The order of the list is important: the benefit of the i^th aggregate is its benefit over the previous aggregates 0, ... i - 1.

This method is not thread safe; you must not call it while calling another method on this Algorithm object. The Algorithm object may or may not have been previously used to generate a set of aggregates (see Algorithm.run(org.pentaho.aggdes.model.Schema, java.util.Map<org.pentaho.aggdes.model.Parameter, java.lang.Object>, org.pentaho.aggdes.algorithm.Progress)).

Parameters:

schema - Schema

parameterValues - Parameter values

aggregateList - List of aggregates

Returns:

list of cost/benefit for each aggregate in the list