Package jline.solvers.ldes

Class LDESResult

java.lang.Object
jline.solvers.SolverResult
jline.solvers.ldes.LDESResult
public class LDESResult extends SolverResult
Since:
1.0
Author:
QORE Lab, Imperial College London
See Also:

  • Field Details

    • tranSysState

      public Map<Integer,Matrix> tranSysState
      Transient system state matrices indexed by time points

    • tranSync

      public Matrix tranSync
      Transient synchronization matrix for multi-class coordination

    • sn

      public NetworkStruct sn
      Network structure used for the analysis

    • QNCI

      public Matrix QNCI
      Confidence interval half-widths for queue lengths [stations x classes]

    • UNCI

      public Matrix UNCI
      Confidence interval half-widths for utilizations [stations x classes]

    • RNCI

      public Matrix RNCI
      Confidence interval half-widths for response times [stations x classes]

    • TNCI

      public Matrix TNCI
      Confidence interval half-widths for throughputs [stations x classes]

    • ANCI

      public Matrix ANCI
      Confidence interval half-widths for arrival rates [stations x classes]

    • WNCI

      public Matrix WNCI
      Confidence interval half-widths for residence times [stations x classes]

    • varianceReductionMethod

      public String varianceReductionMethod
      Variance reduction method used during simulation. Values: "none" (default), "antithetic", "control", "both" - "none": Standard simulation without variance reduction - "antithetic": Antithetic variates using synchronized 1-U method - "control": Control variates using mean-based correction - "both": Combined antithetic and control variates

    • converged

      public boolean converged
      Flag indicating whether the simulation converged before reaching max events/time

    • stoppingReason

      public String stoppingReason
      Stopping reason: "convergence", "max_events" (service-completion budget, options.samples/events, reached), "max_sim_events" (all-event cap, LDESOptions.maxSimEvents, reached), or "max_time" (wall-clock budget).

    • convergenceBatches

      public int convergenceBatches
      Number of batches used for final confidence interval estimation

    • totalSimulatedEvents

      public long totalSimulatedEvents
      Total number of simulated service completion events

    • QNSamples

      public Matrix QNSamples
      Number of samples (observations) used for queue length estimation [stations x classes]

    • UNSamples

      public Matrix UNSamples
      Number of samples (observations) used for utilization estimation [stations x classes]

    • RNSamples

      public Matrix RNSamples
      Number of samples (observations) used for response time estimation [stations x classes]

    • TNSamples

      public Matrix TNSamples
      Number of samples (observations) used for throughput estimation [stations x classes]

    • QNRelPrec

      public Matrix QNRelPrec
      Relative precision achieved for queue lengths (CI half-width / mean) [stations x classes]

    • UNRelPrec

      public Matrix UNRelPrec
      Relative precision achieved for utilizations (CI half-width / mean) [stations x classes]

    • RNRelPrec

      public Matrix RNRelPrec
      Relative precision achieved for response times (CI half-width / mean) [stations x classes]

    • TNRelPrec

      public Matrix TNRelPrec
      Relative precision achieved for throughputs (CI half-width / mean) [stations x classes]

    • renegedCustomers

      public Matrix renegedCustomers
      Number of customers who reneged (abandoned queue due to expired patience) [stations x classes]

    • avgRenegingWaitTime

      public Matrix avgRenegingWaitTime
      Average wait time before reneging [stations x classes]

    • renegingRate

      public Matrix renegingRate
      Reneging rate (reneged / (completed + reneged + dropped)) [stations x classes]

    • balkedCustomers

      public Matrix balkedCustomers
      Number of customers who balked (refused to join upon arrival) [stations x classes]

    • balkingProbability

      public Matrix balkingProbability
      Balking probability (balked / arrivals) [stations x classes]

    • retriedCustomers

      public Matrix retriedCustomers
      Number of successful retrial attempts (customers who re-entered queue from orbit) [stations x classes]

    • retrialDropped

      public Matrix retrialDropped
      Number of customers dropped after exceeding max retrial attempts [stations x classes]

    • avgOrbitSize

      public Matrix avgOrbitSize
      Average orbit size (time-weighted mean jobs in orbit) [stations x classes]

    • respTimeSamples

      public List<Double>[][] respTimeSamples
      Response time samples for empirical CDF computation [stations][classes]. Each element is a list of individual response time observations.

    • passTimeSamples

      public List<Double>[][] passTimeSamples
      Passage time samples for empirical CDF computation [stations][classes]. Passage times track time from arrival to departure including routing delays.

    • tranProbT

      public Matrix tranProbT
      Time points for transient probability analysis.

    • tranProbPit

      public Matrix tranProbPit
      Transient state probabilities over time [numTimePoints x numStates]. Each row contains probability distribution at that time point.

    • tranProbStateSpace

      public Matrix tranProbStateSpace
      Transient state space matrix [numStates x stateVectorLength]. Defines the state corresponding to each column in tranProbPit.

    • tranProbAggrPit

      public Map<Integer,Matrix> tranProbAggrPit
      Transient aggregated state probabilities per node [nodeIdx] -> [numTimePoints x numAggrStates].

    • tranProbAggrStateSpace

      public Map<Integer,Matrix> tranProbAggrStateSpace
      Transient aggregated state space per node [nodeIdx] -> [numAggrStates x numClasses].

    • avgReward

      public Map<String,Double> avgReward
      Steady-state expected Markov reward per reward name, E[r] = (1/T) integral r(X(t)) dt. Populated only when the model defines rewards via model.setReward.

    • tranReward

      public Map<String,double[]> tranReward
      Single-run transient reward trajectory r(X(t)) per reward name.

    • rewardTime

      public Matrix rewardTime
      Time points associated with tranReward.

    • rewardNames

      public List<String> rewardNames
      Names of the reward functions, in registration order.

    • stateHistogramSpace

      public Matrix stateHistogramSpace
      Exact joint-state residence-time histogram (populated when LDESOptions.exportStateHistogram is set). Each row is an aggregated system state in the CTMC stateSpaceAggr layout (station-major, class-minor, nstations*nclasses columns).

    • stateHistogramTime

      public Matrix stateHistogramTime
      Residence time per state in stateHistogramSpace (column vector). Divide by its sum to obtain steady-state state probabilities.

    • stateTrajectorySpace

      public Matrix stateTrajectorySpace
      Integer joint-state trajectory (one aggregated state per row, same layout as stateHistogramSpace) sampled along the simulated path, downsampled to a bounded number of points. Used by MATLAB/native Python for getTranReward.

    • stateTrajectoryTime

      public Matrix stateTrajectoryTime
      Time points associated with stateTrajectorySpace (column vector).

    • busyPeriodMean

      public Matrix busyPeriodMean
      Mean busy period duration per target (rows) and order (columns), populated when LDESOptions.busyPeriodOrders is positive. A busy period of order n runs from the instant an arrival raises the jobs held by the target to n up to the instant the target falls back below n. Row t column n-1 is zero when no period of that order completed.

    • busyPeriodCount

      public Matrix busyPeriodCount
      Number of completed busy periods per target (rows) and order (columns).

    • busyPeriodNames

      public List<String> busyPeriodNames
      Target names, one per row of busyPeriodMean: the station name for an aggregate target and "station:class" for a per-class one.

    • busyPeriodStations

      public List<int[]> busyPeriodStations
      Station indexes forming each target, one entry per row.

    • busyPeriodClass

      public int[] busyPeriodClass
      Class of each target, -1 when the target aggregates every class.

  • Constructor Details

    • LDESResult

      public LDESResult()
      Constructs an empty LDESResult to allow field population from different sources.

    • LDESResult

      public LDESResult(Matrix QN, Matrix UN, Matrix RN, Matrix TN, Matrix CN, Matrix XN, Map<Integer,Matrix> tranSysState, Matrix tranSync, NetworkStruct sn)
      Constructs a LDESResult with the specified performance metrics and state information.
      Parameters:
      QN - queue lengths [stations x classes]
      UN - utilizations [stations x classes]
      RN - response times [stations x classes]
      TN - throughputs [stations x classes]
      CN - visit counts [stations x classes]
      XN - arrival rates [stations x classes]
      tranSysState - transient system states indexed by time
      tranSync - transient synchronization matrix
      sn - network structure information

  • Method Details

    • initRespTimeSamples

      public void initRespTimeSamples(int nStations, int nClasses)
      Initializes the respTimeSamples array with empty lists for each (station, class) pair.
      Parameters:
      nStations - number of stations
      nClasses - number of classes