KSLCore/ksl.simopt.solvers.algorithms/CrossEntropySolver

CrossEntropySolver

class CrossEntropySolver @JvmOverloads constructor(problemDefinition: ProblemDefinition, evaluator: EvaluatorIfc, val ceSampler: CESamplerIfc = CENormalSampler(problemDefinition), maxIterations: Int = ceDefaultMaxIterations, replicationsPerEvaluation: ReplicationPerEvaluationIfc, solutionEqualityChecker: SolutionEqualityIfc = InputsAndConfidenceIntervalEquality(), name: String? = null) : StochasticSolver(source)

Constructs an instance of CrossEntropySolver with specified parameters.

Parameters

evaluator

The evaluator responsible for assessing the quality of solutions. Must implement the EvaluatorIfc interface.

ceSampler

the cross-entropy sampler for the cross-entropy distribution

maxIterations

The maximum number of iterations allowed for the search process.

replicationsPerEvaluation

Strategy to determine the number of replications to perform for each evaluation.

solutionEqualityChecker

Used when testing if solutions have converged for equality between solutions. The default is InputsAndConfidenceIntervalEquality, which checks if the inputs are the same and their is no statistical difference between the solutions

name

Optional name identifier for this instance of solver.

Constructors

CrossEntropySolver

@JvmOverloads

constructor(problemDefinition: ProblemDefinition, evaluator: EvaluatorIfc, ceSampler: CESamplerIfc = CENormalSampler(problemDefinition), maxIterations: Int = ceDefaultMaxIterations, replicationsPerEvaluation: Int = defaultReplicationsPerEvaluation, solutionEqualityChecker: SolutionEqualityIfc = InputsAndConfidenceIntervalEquality(), name: String? = null)

Constructs an instance of CrossEntropySolver with specified parameters.

@JvmOverloads

constructor(problemDefinition: ProblemDefinition, evaluator: EvaluatorIfc, ceSampler: CESamplerIfc = CENormalSampler(problemDefinition), maxIterations: Int = ceDefaultMaxIterations, replicationsPerEvaluation: ReplicationPerEvaluationIfc, solutionEqualityChecker: SolutionEqualityIfc = InputsAndConfidenceIntervalEquality(), name: String? = null)

Types

Companion

object Companion

Properties

ceSampler

val ceSampler: CESamplerIfc

ceSampleSize

var ceSampleSize: Int

The sample size associated with the CE algorithm used to determine the elite solutions. By default, this is determined by the function recommendCESampleSize() within the companion object. The value cannot be less than defaultMinCESampleSize or greater than defaultMaxCESampleSize

elitePct

var elitePct: Double

A value between 0 and 1 that represents the proportion of the CE sample that determines the elite sample. By default, this is 0.1, which can be controlled globally via the companion object's defaultElitePct.

elites

val elites: List<Solution>

The current list of elite solutions ordered from the best

eliteSizeFn

var eliteSizeFn: EliteSizeIfc?

If supplied, this function will be used to determine the size of the elite sample during the cross-entropy process. Supplying a function can permit dynamic changes when determining the elite sample.

sampleSizeFn

var sampleSizeFn: SampleSizeFnIfc?

If supplied, this function will be used to determine the size of the cross-entropy sample during the cross-entropy process. Supplying a function can permit dynamic changes when determining the size of the cross-entropy sample (population).

solutionChecker

val solutionChecker: SolutionChecker

Used to check if the last set of solutions that were captured are the same.

Functions

eliteSize

fun eliteSize(): Int

If eliteSizeFn is supplied it will be used; otherwise, the elite percentage is used to determine the size of the elite sample.

sampleSize

fun sampleSize(): Int

If sampleSizeFn is supplied it will be used; otherwise, the value of ceSampleSize is used to determine the size of the cross-entropy sample (population).

toString

open override fun toString(): String