KSLCore/ksl.utilities.moda

Package-level declarations

Types

class AdditiveMODAModel(metricDefinitions: Map<MetricIfc, ValueFunctionIfc>, weights: Map<MetricIfc, Double> = makeEqualWeights(metricDefinitions.keys), name: String? = null) : MODAModel

Represents a multi-objective decision analysis (MODA) model that uses an additive model for the attribute valuation. The supplied weights must correspond to weights within the model.

AggregationMethod

enum AggregationMethod : Enum<AggregationMethod>

How the alternatives are compared to arrive at a single recommendation.

AlternativeDefinitionResult

data class AlternativeDefinitionResult(val accepted: List<String>, val rejected: List<AlternativeRejection>)

Which of the offered alternatives a model took in, and why it left out any that it did not.

AlternativeRankFrequencyData

data class AlternativeRankFrequencyData(var id: Int = nextId(), var modaName: String = "", var alternative: String = "", var value: Int = 0, var count: Double = 0.0, var proportion: Double = 0.0, var cumProportion: Double = 0.0) : DbTableData

AlternativeRejection

sealed interface AlternativeRejection

The reason an alternative offered to a model was not taken into it.

ElicitationRecord

data class ElicitationRecord(val order: List<String>, val ratings: Map<String, Double>, val elicitedAgainst: Map<String, ElicitedRange>, val adjustableRanges: List<String>)

What was asked, and what was answered, when weights were elicited by the swing method, together with the ranges the answers were given against.

ElicitedRange

data class ElicitedRange(val lowerLimit: Double, val upperLimit: Double)

The range a metric was measured over when a weight was elicited for it.

FlipPoint

data class FlipPoint(val metric: String, val currentWeight: Double, val criticalWeight: Double, val displacedBy: String)

A weight at which the alternative a study currently recommends would be displaced by another.

LinearValueFunction

class LinearValueFunction : ValueFunctionIfc

Uses a linear transformation to transform from score with a metric domain to the value domain.

LogisticFunction

class LogisticFunction(var location: Double = 0.0, scale: Double = 1.0) : ValueFunctionIfc

Metric

open class Metric(val name: String, val domain: Interval = Interval(0.0, Double.MAX_VALUE), val allowLowerLimitAdjustment: Boolean = true, val allowUpperLimitAdjustment: Boolean = true) : MetricIfc

This class serves as a base class for classes that implement the MetricIfc interface. A metric is figure of merit that characterizes the performance of a device, system, method, or entity, relative to its alternatives.

MetricData

data class MetricData(var id: Int = nextId(), var modaName: String = "", var metricName: String = "", var direction: String = "", var weight: Double = 1.0, var domainLowerLimit: Double = 0.0, var domainUpperLimit: Double = Double.POSITIVE_INFINITY, var unitsOfMeasure: String? = null, var description: String? = null, var allowLowerLimitAdjustment: Boolean = false, var allowUpperLimitAdjustment: Boolean = false) : DbTableData

MetricIfc

interface MetricIfc

A performance metric is figure of merit that characterizes the performance of a device, system, method, or entity, relative to its alternatives.

MetricRecord

data class MetricRecord(val name: String, val direction: String, val weight: Double, val declaredLowerLimit: Double, val declaredUpperLimit: Double, val effectiveLowerLimit: Double, val effectiveUpperLimit: Double, val domainWasRescaled: Boolean, val allowLowerLimitAdjustment: Boolean, val allowUpperLimitAdjustment: Boolean, val realizedLowestScore: Double? = null, val realizedHighestScore: Double? = null, val hadTiedScores: Boolean, val valueFunctionId: String, val unitsOfMeasure: String?, val description: String?)

What was known about one metric at the moment a study was evaluated.

MODAAnalyzer

class MODAAnalyzer(alternativeNames: Set<String>, responseDefinitions: Set<MODAAnalyzerData>, responseData: List<WithinRepViewData>)

MODAAnalyzerData

data class MODAAnalyzerData(val responseName: String, val direction: MetricIfc.Direction = MetricIfc.Direction.SmallerIsBetter, var weight: Double = 1.0, val valueFunction: ValueFunctionIfc = LinearValueFunction(), val domain: Interval = Interval(0.0, Double.MAX_VALUE), var unitsOfMeasure: String? = null, var description: String? = null)

MODAModel

abstract class MODAModel(metricDefinitions: Map<MetricIfc, ValueFunctionIfc>, name: String? = null) : IdentityIfc

Defines a base class for creating multi-objective decision analysis (MODA) models.

MODAReportData

interface MODAReportData

The plain, name-keyed data an AdditiveMODAModel report section needs.

ModaSensitivity

class ModaSensitivity(snapshot: ModaSnapshot)

Asks how much a study's recommendation depends on the weights it was given.

ModaSnapshot

data class ModaSnapshot(val name: String, val metrics: List<MetricRecord>, val alternatives: List<String>, val scores: Map<String, Map<String, Double>>, val values: Map<String, Map<String, Double>>, val overallValues: Map<String, Double>, val firstRankCounts: Map<String, Int>, val averageRankings: Map<String, Double>, val rankingMethod: String, val aggregationMethod: AggregationMethod, val primaryRecommendation: String, val warnings: List<ModaWarning>, val replicationAggregation: String? = null)

The result of evaluating a MODA study, complete and standing on its own.

ModaWarning

sealed interface ModaWarning

Something worth telling the user about that came up while a model was being set up or its metric domains adjusted.

MultiAttributeValueFunctionIfc

interface MultiAttributeValueFunctionIfc

OverallValueData

data class OverallValueData(var id: Int = nextId(), var modaName: String = "", var alternative: String = "", var weightedValue: Double = 0.0, var firstRankCount: Int = 0, var averageRank: Double = 0.0) : DbTableData

Score

data class Score(val metric: MetricIfc, val value: Double, var valid: Boolean = true)

A score represents an evaluation of an alternative, system, or entity based on some metric. Each score is related to a metric and has a value that represents the value of the metric. If there is an issue with computing the score, then the property valid indicates whether the score can be trusted (true) or not (false). The default value of the valid property is true. The supplied value must be within the specified domain of the supplied metric; otherwise, an illegal argument exception will occur.

ScoreData

data class ScoreData(var id: Int = nextId(), var modaName: String = "", var alternative: String = "", var scoreName: String = "", var scoreValue: Double = 0.0) : DbTableData

SweepPoint

data class SweepPoint(val weight: Double, val overallValues: Map<String, Double>, val winner: String)

What a study would conclude if one metric carried a given weight.

ValueData

data class ValueData(var id: Int = nextId(), var modaName: String = "", var alternative: String = "", var metricName: String = "", var metricValue: Double = 0.0, var rank: Double = 0.0) : DbTableData

ValueFunction

abstract class ValueFunction : ValueFunctionIfc

A value function maps values from some metric domain to the value range of 0.0, 1.0, where 0.0 implies no value and 1.0 implies maximal value.

ValueFunctionIfc

fun interface ValueFunctionIfc

A value function maps values from some domain to the value range of 0.0, 1.0, where 0.0 implies no value and 1.0 implies maximal value.

Functions

main

fun main()

modaReportData

fun AdditiveMODAModel.modaReportData(): MODAReportData

Projects this model's results into a plain MODAReportData holder. The ranking projections use this model's defaultRankingMethod, matching the behavior of the no-argument ranking accessors, so the rendered report is unchanged from rendering the live model directly.