Package net.simplace.sim.control.generators

Class EquidistantCalSimGenerator

java.lang.Object
net.simplace.sim.control.FWSimulationControlContainer
net.simplace.sim.control.FWSimulationGenerator
net.simplace.sim.control.generators.DefaultSimulationGenerator
net.simplace.sim.control.generators.EquidistantCalSimGenerator
All Implemented Interfaces:
FWSimFieldContainer
public class EquidistantCalSimGenerator extends DefaultSimulationGenerator
Creates equidistant parameter values within a range.

The generator generates for each given parameter count many steps between min and max and creates for each parameter set a simulation.

An optional constraint prohibits the creation of simulations with invalid values that do not match the constraint rule.

Example

|  <generator id="Generator" class="net.simplace.sim.control.generators.EquidistantCalSimGenerator">
|    <cal id="TsumAnthesis" datatype="DOUBLE" target="vTempSumAnthesis">
|      <input id="min" datatype="DOUBLE">0</input>
|      <input id="max" datatype="DOUBLE">1500</input>
|      <input id="count" datatype="INT">31</input>
|    </cal>
|    <cal id="TsumMaturity" datatype="DOUBLE" target="vTempSumMaturity">
|      <input id="min" datatype="DOUBLE">1000</input>
|      <input id="max" datatype="DOUBLE">2000</input>
|      <input id="count" datatype="INT">21</input>
|    </cal>
|    <constraint rule="${Generator.TsumMaturity} > ${Generator.TsumAnthesis}+100">
|    <identifier rule="TSumAnthesis(${Generator.TsumAnthesis})_TSumMaturity(${Generator.TsumMaturity})"/>
|  </generator>

Notice

The number of generated simulation can be considerably high if you use many parameters, as the the parameters counts multiply. E. g. if you have 10 parameters and 50 counts the generator would generate \(50 \cdot 50 \cdot ...\cdot 50 = 50^{10} \approx 10^{17} \) parameter sets.

To illustrate this magnitude: just to store each set of 10 parameters we need 10 bytes. Storing all parameters, we need \(10^{18}\) bytes - or one million harddisks with 1TB capacity. And if a super fast computer would calculate hundred millions of simulations per second, we would still need \(10^9\) seconds or more than 30 years to calculate them.

Author:
Andreas Enders