Package net.simplace.sim.components.util

Class FunctionValueFromTable

java.lang.Object
net.simplace.sim.model.FWSimComponent
net.simplace.sim.components.util.FunctionValueFromTable
All Implemented Interfaces:
net.simplace.sim.util.FWSimFieldContainer
public class FunctionValueFromTable extends net.simplace.sim.model.FWSimComponent
Calculate values of a function defined by tables.

Description

A function y=f(x) can be defined by giving tables for the x and y values.

This component takes as parameters the x-table and y-table as well as the x value. It outputs the calculated y value for the given x value.

The types of x and y can be DOUBLE, INT or CHAR. Depending on the types, there are different options to calculate y.

  • If x is CHAR, then a lookup is performed
  • If x is numeric (DOUBLE or INT) and y is CHAR, then beside lookup also constant interpolation is available.
  • If x and y are numeric, then beside lookup and constant interpolation there is also linear interpolation available.

There are additional options to configure the component

  • if extrapolation should be performed (numeric x)
  • if the last position in the x-table should be cached (improves speed if x values vary slowly)
  • if y is numeric, then a scaling factor can be given

More details

We consider functions

\[ f : \mathbb{X} \rightarrow \mathbb{Y} \]

where the domain $\mathbb{X}$ and range/image $\mathbb{Y}$ can be intervals of real numbers $I \subseteq \mathbb{R}$ or integer numbers $I\subseteq\mathbb{Z}$ or a finit sets $S=\{a_i \in \mathcal{A} | i=1 .. n \in \mathbb{N}\}$.

Technically real numbers are represented by Double values (DOUBLE in the framework context), integer numbers by Integer (INT) and the set elements $a_i$ can be real numbers ($a_i \in \mathbb{R}$), integers ($a_i \in \mathbb{Z}$) or Strings ('CHAR').

The functions are defined by two tables $T_X = {x_i \in \mathbb{X} | i = 1..n}$ and $T_Y = {y_i \in \mathbb{Y} | i = 1..n}$.

If the domain $\mathbb{X}$ is a finite set, then

\[ f(x_i):=y_i \quad\quad \text{(interpolation method NONE)} \]

If the domain $\mathbb{X}$ is a real or integer interval and the range is a finite set, then there are three different options to define $f$.

\[ f(x) := y_{i-1} \text{ if } x_{i-1} \lt x \leq x_{i} \quad\quad \text{(CONSTANT_LEFT)} \] \[ f(x) := \begin{cases}{y_{i-1} \text{ if } \frac{x_i - x_{i-1}}{2} \leq x \lt x_{i}\\y_{i} \text{ if } x_i \leq x \lt \frac{x_{i+1}-x_i}{2}}\end{cases} \quad\quad \text{(CONSTANT_MID)} \] \[ f(x) := y_{i} \text{ if } x_{i} \leq x \lt x_{i+1} \quad\quad \text{(CONSTANT_RIGHT)} \]

If both domain and range are intervals, then the values are linearly interpolated

\[ f(x) := y_i + \frac{y_{i+1}-y_i}{x_{i+1}-x_i}(x_{i+1}-x) \text{ if } x_i \leq x \lt x_{i+1} \quad\quad \text{(LINEAR)} \]

If the user choses to extrapolate the function for values outside the interval $[x_1,x_n]$ then $f(x):=y_1$ for $x \lt x_1$ and $f(x) := y_n$ for $x_n \lt x$.

Notice: as finite sets can consist of number (real or integer) too, one can't infer from the datatype if it should be a finite set or a numeric interval. Therefore the user has to give the interpolation method explicitely.

Configuration, inputs

The sim component needs exactly one of the tables cXValuesDouble (DOUBLEARRAY), cXValuesInt (INTARRAY) or cXValuesChar (CHARARRAY).

It needs at least one of cYValuesDouble (DOUBLEARRAY), cYValuesInt (INTARRAY), cYValuesChar (CHARARRAY).

It needs exaxtly one of the variables iXDouble, iXInt or iXChar. It outputs then at least one of the calculated YDouble, YInt or YChar (depending which of the y-tables are given).

Interpolation method can be defined by setting cMethod to one of NONE, CONSTANT_LEFT,`CONSTANT_MID`, CONSTANT_RIGHT or LINEAR. Setting cExtrapolate to true/`false` will en-/disable extrapolation. If disabled, null will be returned for values outside the range.

cUseCaching determines whether the index in the x-table of the previous function evaluation is stored. If not stored, then the routine will always start at the beginning of the x-table until it reaches the values it sits inbetween. If the x values vary slowly from day to day (e.g. DVS), then storing the last index will improve the speed.

Author:
Gunther Krauss, gk@uni-bonn.de
See Also:

Component Variables

Content TypeNameDescriptionData TypeUnitMin ValueMax ValueDefault Value
constantcExtrapolateIf true, then numeric x values outside the range will give the first (x smaller) or the last (x bigger) y valueBOOLEAN--true
constantcMethodInterpolation method, one of 'NONE' (only lookup), 'CONSTANT_LEFT', 'CONSTANT_MID', 'CONSTANT_RIGHT' or 'LINEAR' CHAR--
constantcScaleValueValue that will be multiplied with the function valueDOUBLE--1.0
constantcUseCachingIf the index of the previous calculation should be stored. Can improve speed, if x values vary slowly.BOOLEAN--true
constantcXValuesCharX-Values of the function table - strings, categorical valuesCHARARRAY---
constantcXValuesDoubleX-Values of the function table - real numbersDOUBLEARRAY---
constantcXValuesIntX-Values of the function table - integer numbersINTARRAY---
constantcYValuesCharY-Values of the function table - strings, categorical valuesCHARARRAY---
constantcYValuesDoubleY-Values of the function table - real numbersDOUBLEARRAY---
constantcYValuesIntY-Values of the function table - integer numbersINTARRAY---
inputiXCharx value (string/categorical) value to look upCHAR--
inputiXDoublex value (real) to calculate y=f(x)DOUBLE--0.0
inputiXIntx value (integer) value to calculateINT--0
outYCharCalculated string/categorical value at iXCHAR--
outYDoubleCalculated real value at iXDOUBLE---
outYIntCalculated integer value at iXINT---

  • Nested Class Summary

    Nested classes/interfaces inherited from class net.simplace.sim.model.FWSimComponent

    net.simplace.sim.model.FWSimComponent.TEST_STATE

  • Field Summary

    Fields inherited from class net.simplace.sim.model.FWSimComponent

    iFieldMap, iFrequence, iInputMap, iJexlRule, iMasterComponentGroup, iName, iOrderNumber, isComponentGroup, iSimComponentElement, iSimModel, iVarMap

  • Constructor Summary

    Constructors
    Constructor
    Description
    FunctionValueFromTable()
    Empty constructor used by class.forName()

  • Method Summary

    Modifier and Type
    Method
    Description
    protected net.simplace.sim.model.FWSimComponent
    clone(net.simplace.sim.util.FWSimVarMap aVarMap)
     
    HashMap<String,net.simplace.sim.util.FWSimVariable<?>>
    createVariables()
    Create the FWSimVariables as interface for this SimComponent
    HashMap<String,net.simplace.sim.util.FWSimVariable<?>>
    fillTestVariables(int aParamIndex, net.simplace.sim.model.FWSimComponent.TEST_STATE aDefineOrCheck)
    called for single component test to check the components algorithm.
    protected void
    init()
    Initializes the fields by getting input and output FWSimVariables from VarMap
    protected void
    process()
    Process the algorithm and write the results back to VarMap

    Methods inherited from class net.simplace.sim.model.FWSimComponent

    addVariable, bind, checkCondition, createSimComponent, createSimComponent, createSimComponent, createSimComponent, doProcess, getConstantVariables, getContentType, getCreateFormXML, getDescription, getEditFormXML, getFieldMap, getFrequence, getFrequenceRuleScript, getInputs, getInputVariables, getMasterComponentGroup, getName, getOrderNumber, getOutputVariables, getVariable, getVariableField, getVarMap, initialize, isConditionCheck, isVariableAvailable, performLinks, performLinks, readInputs, removeVariable, reset, runComponentTest, setVariablesDefault, toComponentLinkingXML, toDocXML, toGroupXML, toOutputDefinitionXML, toResourcesDataXML, toResourcesDefinitionXML, toString, toXML, writeVarInfos

    Methods inherited from class java.lang.Object

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

  • Constructor Details

    • FunctionValueFromTable

      public FunctionValueFromTable()
      Empty constructor used by class.forName()

  • Method Details

    • createVariables

      public HashMap<String,net.simplace.sim.util.FWSimVariable<?>> createVariables()
      Create the FWSimVariables as interface for this SimComponent
      Specified by:
      createVariables in interface net.simplace.sim.util.FWSimFieldContainer
      Specified by:
      createVariables in class net.simplace.sim.model.FWSimComponent
      See Also:
      • FWSimComponent.createVariables()

    • init

      protected void init()
      Initializes the fields by getting input and output FWSimVariables from VarMap
      Specified by:
      init in class net.simplace.sim.model.FWSimComponent
      See Also:
      • FWSimComponent.init()

    • process

      protected void process()
      Process the algorithm and write the results back to VarMap
      Specified by:
      process in class net.simplace.sim.model.FWSimComponent
      See Also:
      • FWSimComponent.process()

    • fillTestVariables

      public HashMap<String,net.simplace.sim.util.FWSimVariable<?>> fillTestVariables(int aParamIndex, net.simplace.sim.model.FWSimComponent.TEST_STATE aDefineOrCheck)
      called for single component test to check the components algorithm.
      Specified by:
      fillTestVariables in class net.simplace.sim.model.FWSimComponent
      See Also:
      • net.simplace.sim.util.FWSimFieldContainer#fillTestVariables(int aParamIndex, TEST_STATE aDefineOrCheck)

    • clone

      protected net.simplace.sim.model.FWSimComponent clone(net.simplace.sim.util.FWSimVarMap aVarMap)
      Specified by:
      clone in class net.simplace.sim.model.FWSimComponent
      See Also:
      • FWSimComponent.clone(net.simplace.sim.util.FWSimVarMap)