Reasoner
Class SimulationEngine

java.lang.Object
  |
  +--Reasoner.SimulationEngine

public class SimulationEngine

extends java.lang.Object

Provides the framework for running the simulation that was set up by all the previous processing stages.

Version:

Created 5 September 2003

Author:

Dan Tappan

See Also:

StaticReasoner, SimulationEngine, DynamicReasoner, MonteCarloRunSet, ObjectTemplateSet

Field Summary

private java.util.ArrayList	m_monte_carlo_run_sets
private ObjectTemplateSet	m_object_template_set
private java.lang.String	m_simulation_name
private StaticReasoner	m_static_reasoner

Constructor Summary

SimulationEngine(java.lang.String simulation_name, StaticReasoner static_reasoner)
Creates a new simulator.

Method Summary

private Dimensions	computeDimensions(ObjectTemplate object_template, boolean force_determinism) Returns the randomly selected dimensions based on the definition of their plausibility intervals and associated attribute interval node in an object template.
private double	computeNondeterministicDimensionValue(PlausibilityInterval plausibility_interval, boolean force_determinism) Returns a randomly selected dimension value based on the definition in a plausibility interval and its associated attribute interval node (i.e., center, variance, bandpass).
private Dimensions	computeProportionedDimensions(ObjectTemplate object_template, double height, double width, double depth) Returns the dimensions proportioned such that a change in one dimension affects the other two dimensions.
java.util.ArrayList	getMonteCarloRunSets() Returns the Monte Carlo run sets generated by this simulation.
java.lang.String	getName() Returns the name of this simulation.
ObjectTemplateSet	getObjectTemplateSet() Returns the object template set supporting this simulation engine.
StaticReasoner	getStaticReasoner() Returns the static reasoner supporting this simulation engine.
private void	simulate(boolean force_determinism, int num_simulations, int max_partition_iterations, int max_partition_backtracks, boolean generate_inferences) Runs a Monte Carlo simulation over the object templates to produce sets of interpretations.
void	simulate(int num_simulations, int max_partition_iterations, int max_partition_backtracks, boolean force_determinism, boolean generate_inferences) Runs a Monte Carlo simulation over the object templates to produce sets of interpretations.
void	simulate(int num_simulations, int max_partition_iterations, int max_partition_backtracks, boolean force_determinism, boolean generate_inferences, long random_seed) Runs a Monte Carlo simulation over the object templates to produce sets of interpretations.
java.lang.String	toString() Returns a string representation of this object.

Methods inherited from class java.lang.Object

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

Field Detail

m_simulation_name

private java.lang.String m_simulation_name

m_object_template_set

private ObjectTemplateSet m_object_template_set

m_static_reasoner

private StaticReasoner m_static_reasoner

m_monte_carlo_run_sets

private java.util.ArrayList m_monte_carlo_run_sets

Constructor Detail

SimulationEngine

public SimulationEngine(java.lang.String simulation_name,
                        StaticReasoner static_reasoner)

Creates a new simulator.

Parameters:

simulation_name - - the simulation name

static_reasoner - - this static reasoner contributing to this simulation engine

Method Detail

computeDimensions

private Dimensions computeDimensions(ObjectTemplate object_template,
                                     boolean force_determinism)
                              throws java.lang.Exception

Returns the randomly selected dimensions based on the definition of their plausibility intervals and associated attribute interval node in an object template.

Parameters:

object_template - - the defining object template

force_determinism - - whether the random-number component of the calculation is disabled

Returns:

the dimensions

Throws:

java.lang.Exception - - if the dimensions cannot be computed

See Also:

computeProportionedDimensions(Reasoner.ObjectTemplate, double, double, double), computeNondeterministicDimensionValue(KnowledgeBase.PlausibilityInterval, boolean)

computeNondeterministicDimensionValue

private double computeNondeterministicDimensionValue(PlausibilityInterval plausibility_interval,
                                                     boolean force_determinism)
                                              throws java.lang.Exception

Returns a randomly selected dimension value based on the definition in a plausibility interval and its associated attribute interval node (i.e., center, variance, bandpass). This is normally called for each dimension property (i.e., height, width, and depth).

Parameters:

plausibility_interval - - the plausibility interval defining the value computation

force_determinism - - whether the random-number component of the calculation is disabled

Returns:

the value

Throws:

java.lang.Exception - - if the dimension cannot be computed

See Also:

computeDimensions(Reasoner.ObjectTemplate, boolean), AttributeIntervalNode, PlausibilityInterval.getCenteredRandomValue(double, double, double), PlausibilityInterval.getBoundFixedValue(int)

computeProportionedDimensions

private Dimensions computeProportionedDimensions(ObjectTemplate object_template,
                                                 double height,
                                                 double width,
                                                 double depth)

Returns the dimensions proportioned such that a change in one dimension affects the other two dimensions. The degree of correlation is defined by the last term of the plausibility interval.

Parameters:

object_template - - the defining object template

height - - the unproportioned height

width - - the unproportioned width

depth - - the unproportioned depth

Returns:

the proportioned dimensions

See Also:

computeDimensions(Reasoner.ObjectTemplate, boolean)

getMonteCarloRunSets

public java.util.ArrayList getMonteCarloRunSets()

Returns the Monte Carlo run sets generated by this simulation.

Returns:

the run sets, or null if the simulation has not been run

getName

public java.lang.String getName()

Returns the name of this simulation.

Returns:

the name

getObjectTemplateSet

public ObjectTemplateSet getObjectTemplateSet()

Returns the object template set supporting this simulation engine.

Returns:

the object template set

getStaticReasoner

public StaticReasoner getStaticReasoner()

Returns the static reasoner supporting this simulation engine.

Returns:

the semantic network

simulate

public void simulate(int num_simulations,
                     int max_partition_iterations,
                     int max_partition_backtracks,
                     boolean force_determinism,
                     boolean generate_inferences)
              throws java.lang.Exception

Runs a Monte Carlo simulation over the object templates to produce sets of interpretations. The centralized random-number generator is seeded with the system clock to produce differing results from nondeterministic runs.

Parameters:

num_simulations - - the number of simulations to run

max_partition_iterations - - the maximum number of propagation attempts per partition before backtracking

max_partition_backtracks - - the maximum number of backtracks to earlier partitions

force_determinism - - whether the random-number component of the dimensions calculations is disabled

generate_inferences - - whether inferences should be generated

Throws:

java.lang.Exception - - if the simulation fails

simulate

public void simulate(int num_simulations,
                     int max_partition_iterations,
                     int max_partition_backtracks,
                     boolean force_determinism,
                     boolean generate_inferences,
                     long random_seed)
              throws java.lang.Exception

Runs a Monte Carlo simulation over the object templates to produce sets of interpretations. The centralized random-number generator is seeded with the given value to produce reproducible results from deterministic runs.

Parameters:

num_simulations - - the number of simulations to run

max_partition_iterations - - the maximum number of propagation attempts per partition before backtracking

max_partition_backtracks - - the maximum number of backtracks to earlier partitions

force_determinism - - whether the random-number component of the dimensions calculations is disabled

generate_inferences - - whether inferences should be generated

random_seed - - the random seed

Throws:

java.lang.Exception - - if the simulation fails

simulate

private void simulate(boolean force_determinism,
                      int num_simulations,
                      int max_partition_iterations,
                      int max_partition_backtracks,
                      boolean generate_inferences)
               throws java.lang.Exception

Runs a Monte Carlo simulation over the object templates to produce sets of interpretations.

Parameters:

force_determinism - - whether the random-number component of the dimensions calculations is disabled

num_simulations - - the number of simulations to run

max_partition_iterations - - the maximum number of propagation attempts per partition before backtracking

max_partition_backtracks - - the maximum number of backtracks to earlier partitions

generate_inferences - - whether inferences should be generated

Throws:

java.lang.Exception - - if the simulation fails

toString

public java.lang.String toString()

Returns a string representation of this object.

Overrides:

toString in class java.lang.Object

Returns:

the String representation

See Also:

Support.toPrettyString(java.lang.String)