Class Solver_mam_inap

java.lang.Object
jline.solvers.mam.handlers.Solver_mam_inap

public final class Solver_mam_inap extends Object
  • Method Details

    • solver_mam_inap

      public static INAPResult solver_mam_inap(RCATModel rcat, double tol, int maxiter, int seed, String method)
    • solver_mam_inap

      public static INAPResult solver_mam_inap(RCATModel rcat, double tol, int maxiter, int seed)
    • solver_mam_inap

      public static INAPResult solver_mam_inap(RCATModel rcat, double tol, int maxiter)
    • solver_mam_inap

      public static INAPResult solver_mam_inap(RCATModel rcat)
    • solver_mam_inapinf

      public static INAPResult solver_mam_inapinf(RCATModel rcat, boolean[] isOpenProc, double tol, int maxiter)
      Matrix-geometric INAP ('inapinf'): the isolated OPEN components are solved directly on their infinite state space by a scalar matrix-geometric (QBD / catastrophe) decomposition -- the marginal is geometric pi_n = (1-rho) rho^n with rho the sub-unit root of the QBD characteristic equation, and any catastrophe drain to the empty state folds into the local outflow. Closed components remain finite. The reversed rate is updated by the weighted-mean formula Eq. (4) evaluated in closed form on the geometric tail, and the RCAT product-form residual (Remark 2) is returned. Mirrors solver_mam_ag.m (inap_inf). Reference: A. Marin, S. Rota Bulo, S. Balsamo, "A Numerical Algorithm for the Decomposition of Cooperating Structured Markov Processes", MASCOTS 2012.