Package jline.solvers.fluid

Class FluidMoments

java.lang.Object
jline.solvers.fluid.FluidMoments
public class FluidMoments extends Object
Second-order report of the fluid moment-closure methods ('minnormal', the cache closure of 'rmf', and 'dae'), returned by SolverFluid.getMoments(). Mirrors the struct returned by MATLAB SolverFLD.getMoments and the dict returned by the native python twin.

Index blocks are 0-based. classBlock is nested as classBlock[i][k], as in MATLAB and python, rather than flattened.

  • Nested Class Summary

    Nested Classes
    Modifier and Type
    Class
    Description
    static class 
    FluidMoments.CacheMoments
    Per-cache second moments of the item occupancy under the linear noise approximation.

  • Field Summary

    Fields
    Modifier and Type
    Field
    Description
    List<FluidMoments.CacheMoments>
    cache
    Per-cache occupancy moments; empty unless the closure ran on a cache model.
    int[][][]
    classBlock
    State coordinates of each (station,class) in Sigma, as classBlock[i][k].
    double
    conservation
    'dae' only: largest violation of the population-conservation rows, else NaN.
    boolean
    converged
    'dae' only: whether that residual reached options.tol.
    String
    method
    The method that produced the moments, as recorded on the result.
    int
    outerIters
    Outer (mean, covariance) iterations performed.
    PetriSolver.PetriReport
    petri
    Petri-net report (marking, its variance, flows, invariants); null unless the model is an SPN.
    Matrix[]
    QCovt
    'dae' transient covariance aggregated onto station-class pairs (ir = r*M + i); null otherwise.
    Matrix
    QStd
    Per-(station,class) queue-length standard deviation, sqrt(max(0, QVar)).
    Matrix
    QVar
    Per-(station,class) queue-length variance.
    Matrix[][]
    QVart
    'dae' transient queue-length variance [station][class] over the time grid; null otherwise.
    Matrix
    refinement
    The 1/N correction of 'mnrefined' in state coordinates; null for every other method.
    double
    residual
    'dae' only: infinity norm of the residual the simultaneous solve stopped at, else NaN.
    Matrix
    Sigma
    State-level stationary covariance: the linear noise Lyapunov solution at the fixed point.
    Matrix
    sigma2
    Per-station population variance.
    Matrix
    sigma2Drift
    The variance as it entered the drift, zero at the delay stations.
    Matrix[]
    Sigmat
    'dae' transient covariance, one matrix per point of the mean's time grid; null otherwise.
    int[][]
    stationBlock
    State coordinates of each station in Sigma.

  • Constructor Summary

    Constructors
    Constructor
    Description
    FluidMoments()
     

  • Method Summary

    Methods inherited from class java.lang.Object

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

  • Field Details

    • Sigma

      public Matrix Sigma
      State-level stationary covariance: the linear noise Lyapunov solution at the fixed point.

    • QVar

      public Matrix QVar
      Per-(station,class) queue-length variance.

    • QStd

      public Matrix QStd
      Per-(station,class) queue-length standard deviation, sqrt(max(0, QVar)).

    • sigma2

      public Matrix sigma2
      Per-station population variance.

    • sigma2Drift

      public Matrix sigma2Drift
      The variance as it entered the drift, zero at the delay stations.

    • outerIters

      public int outerIters
      Outer (mean, covariance) iterations performed.

    • refinement

      public Matrix refinement
      The 1/N correction of 'mnrefined' in state coordinates; null for every other method.

    • stationBlock

      public int[][] stationBlock
      State coordinates of each station in Sigma. Null on the Petri route.

    • classBlock

      public int[][][] classBlock
      State coordinates of each (station,class) in Sigma, as classBlock[i][k]. Null on the Petri route.

    • cache

      public List<FluidMoments.CacheMoments> cache
      Per-cache occupancy moments; empty unless the closure ran on a cache model.

    • petri

      public PetriSolver.PetriReport petri
      Petri-net report (marking, its variance, flows, invariants); null unless the model is an SPN.

    • method

      public String method
      The method that produced the moments, as recorded on the result.

    • residual

      public double residual
      'dae' only: infinity norm of the residual the simultaneous solve stopped at, else NaN.

    • converged

      public boolean converged
      'dae' only: whether that residual reached options.tol.

    • conservation

      public double conservation
      'dae' only: largest violation of the population-conservation rows, else NaN.

    • Sigmat

      public Matrix[] Sigmat
      'dae' transient covariance, one matrix per point of the mean's time grid; null otherwise.

    • QVart

      public Matrix[][] QVart
      'dae' transient queue-length variance [station][class] over the time grid; null otherwise.

    • QCovt

      public Matrix[] QCovt
      'dae' transient covariance aggregated onto station-class pairs (ir = r*M + i); null otherwise.

  • Constructor Details

    • FluidMoments

      public FluidMoments()