Package jline.inference.api

Class VariationalSpec

java.lang.Object
jline.inference.api.VariationalSpec
public class VariationalSpec extends Object
Inference problem handed to Infer_variational.

Station-class pairs are flattened column-major, so that pair (m,r) sits at index r*M+m, matching the MATLAB, Python and C++ specifications. Station and class indices inside arcs are one-based; index 0 marks the external source or sink.

  • Field Summary

    Fields
    Modifier and Type
    Field
    Description
    double[]
    alpha0
    (P) Gamma prior shapes.
    int[]
    arcparam
    (T) index in 1..P of the rate governing the transition, 0 when known.
    double[]
    arcrate
    (T) known rate for transitions with arcparam == 0.
    int[][]
    arcs
    (T x 3) transitions [i j c]; i==0 external source, j==0 sink.
    double[]
    beta0
    (P) Gamma prior rates.
    double[]
    capacity
    (M*R) upper bound on the queue length, infinite by default.
    double
    epsilon
    probability that a reading is faulty.
    double[]
    nservers
    (M) number of servers.
    double[][]
    obsData
    (K x M*R) observed queue lengths, NaN where not observed.
    double[]
    obsRange
    (M*R) support size of the uniform contamination.
    double[]
    obsTimes
    (K) observation epochs.
    double[]
    routeprob
    (T) routing probability of each transition.
    int[]
    sched
    (M) discipline codes: 0 = infinite server, 1 = shared server, 2 = external.
    double[][]
    x0
    (M x R) initial queue lengths.

  • Constructor Summary

    Constructors
    Constructor
    Description
    VariationalSpec()
     

  • Method Summary

    Modifier and Type
    Method
    Description
    static double[]
    flatten(double[][] a)
    Flattened (M*R) copy of a station-class array, column-major.
    int
    narcs()
     
    int
    nclasses()
     
    int
    nparams()
     
    int
    nstations()
     

    Methods inherited from class java.lang.Object

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

  • Field Details

    • arcs

      public int[][] arcs
      (T x 3) transitions [i j c]; i==0 external source, j==0 sink.

    • x0

      public double[][] x0
      (M x R) initial queue lengths.

    • sched

      public int[] sched
      (M) discipline codes: 0 = infinite server, 1 = shared server, 2 = external.

    • nservers

      public double[] nservers
      (M) number of servers.

    • routeprob

      public double[] routeprob
      (T) routing probability of each transition.

    • arcparam

      public int[] arcparam
      (T) index in 1..P of the rate governing the transition, 0 when known.

    • arcrate

      public double[] arcrate
      (T) known rate for transitions with arcparam == 0.

    • alpha0

      public double[] alpha0
      (P) Gamma prior shapes.

    • beta0

      public double[] beta0
      (P) Gamma prior rates.

    • obsTimes

      public double[] obsTimes
      (K) observation epochs.

    • obsData

      public double[][] obsData
      (K x M*R) observed queue lengths, NaN where not observed.

    • obsRange

      public double[] obsRange
      (M*R) support size of the uniform contamination.

    • epsilon

      public double epsilon
      probability that a reading is faulty.

    • capacity

      public double[] capacity
      (M*R) upper bound on the queue length, infinite by default. In a closed network this is the chain population, and clamping the load there keeps the expanded state space from crediting a station with more jobs than the network holds.

  • Constructor Details

    • VariationalSpec

      public VariationalSpec()

  • Method Details

    • nstations

      public int nstations()

    • nclasses

      public int nclasses()

    • narcs

      public int narcs()

    • nparams

      public int nparams()

    • flatten

      public static double[] flatten(double[][] a)
      Flattened (M*R) copy of a station-class array, column-major.