Package jline.solvers
Class SolverResult
java.lang.Object
jline.solvers.SolverResult
- Direct Known Subclasses:
AGResult,CTMCResult,FluidResult,JMTResult,LayeredSolverResult,LDESResult,LNLDESResult,MAMResult,MVAResult,NCResult,ResultCTMC,ResultCTMCMargAggr,SSAResult
Container for storing performance metrics computed by queueing network solvers.
This class holds both steady-state and transient performance results including queue lengths, utilizations, response times, throughputs, and arrival rates. Results are organized as matrices with dimensions [stations x classes] for station-level metrics and [1 x classes] for system-level metrics.
For solver-specific results, create separate classes (e.g., SolverFluidResult).
- See Also:
-
Field Summary
FieldsModifier and TypeFieldDescriptionMean arrival rates [stations x classes]FCR mean arrival rates (attempted entries) [regions x classes]Per-cache-node average results, one record per Cache in the model that produced them, or null when the solve reached no cache.Mean system response times [1 x chains]Matrix[][]Transient system response times [time_points][1 x chains]Fork-Join sibling-drop rate [stations x classes]: at a quorum/PARTIAL Join, the rate of forked siblings discarded because they arrive after the quorum already fired.FCR mean drop rate: the rate of jobs refused entry to the region and dropped (DROP rule), per [regions x classes].intNumber of iterations performed by iterative solversFCR mean memory occupation (time-average of sum of per-class classSize times in-region job count), the JMT "FCR Memory Occupation" measure [regions x classes].The solution method used to compute these resultsTransient probability distributions over timeMean queue lengths [stations x classes]FCR mean queue lengths (jobs waiting at ingress) [regions x classes]Matrix[][]Mean response times [stations x classes]FCR mean response times (wait time at ingress) [regions x classes]Matrix[][]Transient response times [time_points][stations x classes]`options.config.runLengthPlan`: the run length the caller would need for the precision they asked for, as computed bySimRunlength.sim_runlength_plan(jline.util.matrix.Matrix, jline.util.matrix.Matrix, double, double, double).doubleComputation time in secondsThe name of the solver that computed these resultsSteady-state probability distributionMean system tardiness [1 x classes]Time points for transient analysisMean tardiness [stations x classes]booleanTrue if the solver stopped early because the wall-clock time budget (SolverOptions.timeout) was exceeded.Mean throughputs [stations x classes]FCR mean throughputs (entry rate) [regions x classes]Matrix[][]Transient throughputs [time_points][stations x classes]Mean server utilizations [stations x classes]FCR mean utilizations (capacity utilization) [regions x classes]Matrix[][]Transient utilizations [time_points][stations x classes]FCR mean total weight (time-average of sum of per-class classWeight times in-region job count), the JMT "FCR Total Weight" measure [regions x classes].Mean residence times [stations x classes]FCR mean residence times [regions x classes]Mean system throughputs [1 x chains]Matrix[][]Transient system throughputs [time_points][1 x chains] -
Constructor Summary
Constructors -
Method Summary
-
Field Details
-
method
The solution method used to compute these results -
solver
The name of the solver that computed these results -
iter
public int iterNumber of iterations performed by iterative solvers -
QN
Mean queue lengths [stations x classes] -
UN
Mean server utilizations [stations x classes] -
RN
Mean response times [stations x classes] -
TN
Mean throughputs [stations x classes] -
AN
Mean arrival rates [stations x classes] -
WN
Mean residence times [stations x classes] -
TardN
Mean tardiness [stations x classes] -
SysTardN
Mean system tardiness [1 x classes] -
CN
Mean system response times [1 x chains] -
XN
Mean system throughputs [1 x chains] -
QNfcr
FCR mean queue lengths (jobs waiting at ingress) [regions x classes] -
UNfcr
FCR mean utilizations (capacity utilization) [regions x classes] -
RNfcr
FCR mean response times (wait time at ingress) [regions x classes] -
TNfcr
FCR mean throughputs (entry rate) [regions x classes] -
ANfcr
FCR mean arrival rates (attempted entries) [regions x classes] -
WNfcr
FCR mean residence times [regions x classes] -
WeightNfcr
FCR mean total weight (time-average of sum of per-class classWeight times in-region job count), the JMT "FCR Total Weight" measure [regions x classes]. Row sum over classes gives the region-aggregate weight. -
MemOccNfcr
FCR mean memory occupation (time-average of sum of per-class classSize times in-region job count), the JMT "FCR Memory Occupation" measure [regions x classes]. Row sum over classes gives the region-aggregate memory. -
DropRateNfcr
FCR mean drop rate: the rate of jobs refused entry to the region and dropped (DROP rule), per [regions x classes]. Zero for BLOCK regions. The region loss ratio is DropRateNfcr / (TNfcr + DropRateNfcr). -
DropRateJoin
Fork-Join sibling-drop rate [stations x classes]: at a quorum/PARTIAL Join, the rate of forked siblings discarded because they arrive after the quorum already fired. Non-zero only at Join station rows; null when the model has no Join or no quorum. Folded into getAvgLossTable as the Join-row LossRate (the station identity ArvR - Tput does not apply at a synchronizing node). -
runLengthPlan
`options.config.runLengthPlan`: the run length the caller would need for the precision they asked for, as computed bySimRunlength.sim_runlength_plan(jline.util.matrix.Matrix, jline.util.matrix.Matrix, double, double, double). Null unless the option was set and the solver produced confidence intervals. -
QNt
-
RNt
Transient response times [time_points][stations x classes] -
UNt
Transient utilizations [time_points][stations x classes] -
TNt
Transient throughputs [time_points][stations x classes] -
XNt
Transient system throughputs [time_points][1 x chains] -
CNt
Transient system response times [time_points][1 x chains] -
t
Time points for transient analysis -
pi_t
Transient probability distributions over time -
SS
Steady-state probability distribution -
runtime
public double runtimeComputation time in seconds -
timedOut
public boolean timedOutTrue if the solver stopped early because the wall-clock time budget (SolverOptions.timeout) was exceeded. When true, the stored metrics are an interim solution (if any iteration completed) or empty (all null). -
cacheResults
Per-cache-node average results, one record per Cache in the model that produced them, or null when the solve reached no cache.A Cache carries its result on the NODE rather than in Q/U/R/T, so a solver that delegates to another model has no way to hand the surface back through the matrices above. This list is that hand-back; see
CacheResultRecordandNetworkSolver.applyCacheResults(java.util.List).
-
-
Constructor Details
-
SolverResult
public SolverResult()
-
-
Method Details
-
deepCopy
Creates a deep copy of this SolverResult instance. All matrix fields are cloned to ensure independence.- Returns:
- a deep copy of this result object
-
reset
public void reset()Resets all stored results to null and runtime to zero.
-