Package jline.examples.java.advanced

Class AgentModelExamples

java.lang.Object
jline.examples.java.advanced.AgentModelExamples
public class AgentModelExamples extends Object
Examples demonstrating SolverAG, LINE's agent-based solver, and its INAP methods. SolverAG decomposes queueing networks into interacting stochastic processes. INAP (Iterative Numerical Approximation Procedure) efficiently solves the resulting fixed-point equations. These methods used to be reached as SolverMAM('inap'); they are SolverAG's since those methods moved there, and SolverMAM no longer lists them. References: - Marin and Rota-Bulo', "A Mean-Field Analysis of a Class of Interactive Distributed Systems", MASCOTS 2009 - Harrison and Llado, "Stochastic bounds and product form solutions using RCAT", ICPE 2011

  • Constructor Details

    • AgentModelExamples

      public AgentModelExamples()

  • Method Details

    • ag_tandem_open

      public static void ag_tandem_open() throws Exception
      Open tandem queue (M/M/1 -> M/M/1) example. Demonstrates SolverAG with the INAP method on a simple open network with two queues in series. Compares results against analytical M/M/1 formulas and MVA.
      Throws:
      Exception - if the solver encounters an error

    • ag_tandem_phasetype

      public static void ag_tandem_phasetype() throws Exception
      Open tandem queue with PHASE-TYPE service. Queue1 sees the Poisson source directly, so it is an isolated M/PH/1 and its mean queue length must be the Pollaczek-Khinchine value whatever the reversed-rate iteration does. Both stations carry the same mean service time and differ only in variability, which the earlier scalar birth-death construction could not see: it returned the M/M/1 answer for either.
      Throws:
      Exception - if the solver encounters an error

    • ag_closed_network

      public static void ag_closed_network() throws Exception
      Closed network with two PS queues. Demonstrates SolverAG with the INAP method on a closed queueing network with processor-sharing discipline. Compares results against MVA and CTMC.
      Throws:
      Exception - if the solver encounters an error

    • ag_multiclass_closed

      public static void ag_multiclass_closed() throws Exception
      Multiclass closed network. Demonstrates SolverAG with the INAP method on a multiclass closed network. The solver creates separate processes for each (station, class) pair.
      Throws:
      Exception - if the solver encounters an error

    • ag_jackson_network

      public static void ag_jackson_network() throws Exception
      Jackson network with probabilistic routing. Demonstrates SolverAG with the INAP method on an open Jackson network with feedback routing. The solver models job transfers as synchronization actions between processes.
      Throws:
      Exception - if the solver encounters an error

    • ag_gnetwork

      public static void ag_gnetwork() throws Exception
      G-network (Gelenbe network) with negative customers. Demonstrates SolverAG with the INAP method on a G-network where negative customers (signals) remove jobs from queues. This models scenarios like job cancellations or service interrupts. Reference: Gelenbe, E. (1991). "Product-form queueing networks with negative and positive customers", Journal of Applied Probability
      Throws:
      Exception - if the solver encounters an error

    • main

      public static void main(String[] args) throws Exception
      Run all agent model examples.
      Parameters:
      args - command line arguments (unused)
      Throws:
      Exception - if any solver encounters an error