Package jline.api.pfqn.mva

Class Pfqn_schmidt_amva

java.lang.Object
jline.api.pfqn.mva.Pfqn_schmidt_amva
public final class Pfqn_schmidt_amva extends Object
Schmidt MVA algorithm for multi-class queueing networks with general scheduling disciplines (INF, PS, FCFS, including class-dependent multi-server FCFS).

Port of matlab/src/api/pfqn/pfqn_schmidt.m and pfqn_schmidt_ext.m. Both take the service DEMAND matrix D (M x R), as the MATLAB functions do; the pure service times needed by the Bcn terms are recovered as D ./ v.

Reference: R. Schmidt, "An approximate MVA algorithm for exponential, class-dependent multiple server stations," Performance Evaluation, vol. 29, no. 4, pp. 245-254, 1997.

  • Method Details

    • pfqn_schmidt

      public static Ret.pfqnAMVASchmidt pfqn_schmidt(Matrix D, Matrix N, Matrix S, List<SchedStrategy> sched)
      Schmidt's exact MVA, default visit ratios (ones).
      Parameters:
      D - service demand matrix (M x R)
      N - population vector (1 x R)
      S - number of servers per station (M x 1, or M x R per class)
      sched - scheduling discipline per station

    • pfqn_schmidt

      public static Ret.pfqnAMVASchmidt pfqn_schmidt(Matrix D, Matrix N, Matrix S, Matrix v, List<SchedStrategy> sched)
      Schmidt's exact MVA for networks with general scheduling disciplines ([XN,QN,UN,CN] = pfqn_schmidt(D,N,S,sched,v)).
      Parameters:
      D - service demand matrix (M x R)
      N - population vector (1 x R)
      S - number of servers per station (M x 1, or M x R per class)
      v - visit ratio matrix (M x R); null for ones
      sched - scheduling discipline per station
      Returns:
      X (1 x R), Q (M x R), U (M x R) = D*X/nservers, C (M x R) residence times

    • pfqn_schmidt_ext

      public static Ret.pfqnAMVASchmidt pfqn_schmidt_ext(Matrix D, Matrix N, Matrix S, List<SchedStrategy> sched)
      Extended Schmidt MVA algorithm with queue-aware alpha corrections ([XN,QN,UN,CN] = pfqn_schmidt_ext(D,N,S,sched)). A queue-aware version of the Schmidt algorithm that precomputes alpha values for improved accuracy in networks with class-dependent FCFS scheduling.
      Parameters:
      D - service demand matrix (M x R)
      N - population vector (1 x R)
      S - number of servers per station (M x 1)
      sched - scheduling discipline per station