[17487] | 1 | #region License Information
|
---|
| 2 | /*
|
---|
| 3 | * This file is part of SimSharp which is licensed under the MIT license.
|
---|
| 4 | * See the LICENSE file in the project root for more information.
|
---|
| 5 | */
|
---|
| 6 | #endregion
|
---|
| 7 |
|
---|
| 8 | using System;
|
---|
| 9 | using System.Collections.Generic;
|
---|
| 10 | using System.Diagnostics;
|
---|
| 11 | using System.IO;
|
---|
| 12 | using System.Threading;
|
---|
| 13 | using System.Threading.Tasks;
|
---|
| 14 |
|
---|
| 15 | namespace SimSharp {
|
---|
| 16 | /// <summary>
|
---|
| 17 | /// Simulation hold the event queues, schedule and process events.
|
---|
| 18 | /// </summary>
|
---|
| 19 | /// <remarks>
|
---|
| 20 | /// This class is not thread-safe against manipulation of the event queue. If you supply a termination
|
---|
| 21 | /// event that is set outside the simulation thread, please use the <see cref="ThreadSafeSimulation"/> environment.
|
---|
| 22 | ///
|
---|
| 23 | /// For most purposes <see cref="Simulation"/> is however the better and faster choice.
|
---|
| 24 | /// </remarks>
|
---|
| 25 | public class Simulation {
|
---|
| 26 | private const int InitialMaxEvents = 1024;
|
---|
| 27 |
|
---|
| 28 | /// <summary>
|
---|
| 29 | /// Describes the number of seconds that a logical step of 1 in the *D-API takes.
|
---|
| 30 | /// </summary>
|
---|
| 31 | protected double DefaultTimeStepSeconds { get; private set; }
|
---|
| 32 |
|
---|
| 33 | /// <summary>
|
---|
| 34 | /// Calculates the logical date of the simulation by the amount of default steps
|
---|
| 35 | /// that have passed.
|
---|
| 36 | /// </summary>
|
---|
| 37 | public double NowD {
|
---|
| 38 | get { return (Now - StartDate).TotalSeconds / DefaultTimeStepSeconds; }
|
---|
| 39 | }
|
---|
| 40 |
|
---|
| 41 | private DateTime now;
|
---|
| 42 | /// <summary>
|
---|
| 43 | /// The current simulation time as a calendar date.
|
---|
| 44 | /// </summary>
|
---|
| 45 | public virtual DateTime Now { get => now; protected set => now = value; }
|
---|
| 46 |
|
---|
| 47 | /// <summary>
|
---|
| 48 | /// The calendar date when the simulation started. This defaults to 1970-1-1 if
|
---|
| 49 | /// no other date has been specified in the overloaded constructor.
|
---|
| 50 | /// </summary>
|
---|
| 51 | public DateTime StartDate { get; protected set; }
|
---|
| 52 |
|
---|
| 53 | /// <summary>
|
---|
| 54 | /// The random number generator that is to be used in all events in
|
---|
| 55 | /// order to produce reproducible results.
|
---|
| 56 | /// </summary>
|
---|
| 57 | protected IRandom Random { get; set; }
|
---|
| 58 |
|
---|
| 59 | protected EventQueue ScheduleQ;
|
---|
| 60 | public Process ActiveProcess { get; set; }
|
---|
| 61 |
|
---|
| 62 | public TextWriter Logger { get; set; }
|
---|
| 63 | public int ProcessedEvents { get; protected set; }
|
---|
| 64 |
|
---|
| 65 | public Simulation() : this(new DateTime(1970, 1, 1)) { }
|
---|
| 66 | public Simulation(TimeSpan? defaultStep) : this(new DateTime(1970, 1, 1), defaultStep) { }
|
---|
| 67 | public Simulation(int randomSeed, TimeSpan? defaultStep = null) : this(new DateTime(1970, 1, 1), randomSeed, defaultStep) { }
|
---|
| 68 | public Simulation(DateTime initialDateTime, TimeSpan? defaultStep = null) : this(new PcgRandom(), initialDateTime, defaultStep) { }
|
---|
| 69 | public Simulation(DateTime initialDateTime, int randomSeed, TimeSpan? defaultStep = null) : this(new PcgRandom(randomSeed), initialDateTime, defaultStep) { }
|
---|
| 70 | public Simulation(IRandom random, DateTime initialDateTime, TimeSpan? defaultStep = null) {
|
---|
| 71 | DefaultTimeStepSeconds = (defaultStep ?? TimeSpan.FromSeconds(1)).Duration().TotalSeconds;
|
---|
| 72 | StartDate = initialDateTime;
|
---|
| 73 | Now = initialDateTime;
|
---|
| 74 | Random = random;
|
---|
| 75 | ScheduleQ = new EventQueue(InitialMaxEvents);
|
---|
| 76 | Logger = Console.Out;
|
---|
| 77 | }
|
---|
| 78 |
|
---|
| 79 | public double ToDouble(TimeSpan span) {
|
---|
| 80 | return span.TotalSeconds / DefaultTimeStepSeconds;
|
---|
| 81 | }
|
---|
| 82 |
|
---|
| 83 | public TimeSpan ToTimeSpan(double span) {
|
---|
| 84 | return TimeSpan.FromSeconds(DefaultTimeStepSeconds * span);
|
---|
| 85 | }
|
---|
| 86 |
|
---|
| 87 | /// <summary>
|
---|
| 88 | /// Creates a new process from an event generator. The process is automatically
|
---|
| 89 | /// scheduled to be started at the current simulation time.
|
---|
| 90 | /// </summary>
|
---|
| 91 | /// <param name="generator">The generator function that represents the process.</param>
|
---|
| 92 | /// <param name="priority">The priority to rank events at the same time (smaller value = higher priority).</param>
|
---|
| 93 | /// <returns>The scheduled process that was created.</returns>
|
---|
| 94 | public Process Process(IEnumerable<Event> generator, int priority = 0) {
|
---|
| 95 | return new Process(this, generator, priority);
|
---|
| 96 | }
|
---|
| 97 |
|
---|
| 98 | /// <summary>
|
---|
| 99 | /// Creates and returns a new timeout.
|
---|
| 100 | /// </summary>
|
---|
| 101 | /// <param name="delay">The time after which the timeout is fired.</param>
|
---|
| 102 | /// <param name="priority">The priority to rank events at the same time (smaller value = higher priority).</param>
|
---|
| 103 | /// <returns>The scheduled timeout event that was created.</returns>
|
---|
| 104 | public Timeout TimeoutD(double delay, int priority = 0) {
|
---|
| 105 | return Timeout(TimeSpan.FromSeconds(DefaultTimeStepSeconds * delay), priority);
|
---|
| 106 | }
|
---|
| 107 |
|
---|
| 108 | /// <summary>
|
---|
| 109 | /// Creates and returns a new timeout.
|
---|
| 110 | /// </summary>
|
---|
| 111 | /// <param name="delay">The time after which the timeout is fired.</param>
|
---|
| 112 | /// <param name="priority">The priority to rank events at the same time (smaller value = higher priority).</param>
|
---|
| 113 | /// <returns>The scheduled timeout event that was created.</returns>
|
---|
| 114 | public Timeout Timeout(TimeSpan delay, int priority = 0) {
|
---|
| 115 | return new Timeout(this, delay, priority: priority);
|
---|
| 116 | }
|
---|
| 117 |
|
---|
| 118 | public virtual void Reset(int randomSeed) {
|
---|
| 119 | ProcessedEvents = 0;
|
---|
| 120 | Now = StartDate;
|
---|
| 121 | Random = new PcgRandom(randomSeed);
|
---|
| 122 | ScheduleQ = new EventQueue(InitialMaxEvents);
|
---|
| 123 | useSpareNormal = false;
|
---|
| 124 | }
|
---|
| 125 |
|
---|
| 126 | public virtual void ScheduleD(double delay, Event @event) {
|
---|
| 127 | Schedule(TimeSpan.FromSeconds(DefaultTimeStepSeconds * delay), @event);
|
---|
| 128 | }
|
---|
| 129 |
|
---|
| 130 | /// <summary>
|
---|
| 131 | /// Schedules an event to occur at the same simulation time as the call was made.
|
---|
| 132 | /// </summary>
|
---|
| 133 | /// <param name="event">The event that should be scheduled.</param>
|
---|
| 134 | /// <param name="priority">The priority to rank events at the same time (smaller value = higher priority).</param>
|
---|
| 135 | public virtual void Schedule(Event @event, int priority = 0) {
|
---|
| 136 | DoSchedule(Now, @event, priority);
|
---|
| 137 | }
|
---|
| 138 |
|
---|
| 139 | /// <summary>
|
---|
| 140 | /// Schedules an event to occur after a certain (positive) delay.
|
---|
| 141 | /// </summary>
|
---|
| 142 | /// <exception cref="ArgumentException">
|
---|
| 143 | /// Thrown when <paramref name="delay"/> is negative.
|
---|
| 144 | /// </exception>
|
---|
| 145 | /// <param name="delay">The (positive) delay after which the event should be fired.</param>
|
---|
| 146 | /// <param name="event">The event that should be scheduled.</param>
|
---|
| 147 | /// <param name="priority">The priority to rank events at the same time (smaller value = higher priority).</param>
|
---|
| 148 | public virtual void Schedule(TimeSpan delay, Event @event, int priority = 0) {
|
---|
| 149 | if (delay < TimeSpan.Zero)
|
---|
| 150 | throw new ArgumentException("Negative delays are not allowed in Schedule(TimeSpan, Event).");
|
---|
| 151 | var eventTime = Now + delay;
|
---|
| 152 | DoSchedule(eventTime, @event, priority);
|
---|
| 153 | }
|
---|
| 154 |
|
---|
| 155 | protected virtual EventQueueNode DoSchedule(DateTime date, Event @event, int priority = 0) {
|
---|
| 156 | if (ScheduleQ.MaxSize == ScheduleQ.Count) {
|
---|
| 157 | // the capacity has to be adjusted, there are more events in the queue than anticipated
|
---|
| 158 | var oldSchedule = ScheduleQ;
|
---|
| 159 | ScheduleQ = new EventQueue(ScheduleQ.MaxSize * 2);
|
---|
| 160 | foreach (var e in oldSchedule) ScheduleQ.Enqueue(e.PrimaryPriority, e.Event, e.SecondaryPriority);
|
---|
| 161 | }
|
---|
| 162 | return ScheduleQ.Enqueue(date, @event, priority);
|
---|
| 163 | }
|
---|
| 164 |
|
---|
| 165 | public virtual object RunD(double? until = null) {
|
---|
| 166 | if (!until.HasValue) return Run();
|
---|
| 167 | return Run(Now + TimeSpan.FromSeconds(DefaultTimeStepSeconds * until.Value));
|
---|
| 168 | }
|
---|
| 169 |
|
---|
| 170 | public virtual object Run(TimeSpan span) {
|
---|
| 171 | return Run(Now + span);
|
---|
| 172 | }
|
---|
| 173 |
|
---|
| 174 | public virtual object Run(DateTime until) {
|
---|
| 175 | if (until <= Now) throw new InvalidOperationException("Simulation end date must lie in the future.");
|
---|
| 176 | var stopEvent = new Event(this);
|
---|
| 177 | var node = DoSchedule(until, stopEvent);
|
---|
| 178 | // stop event is always the first to execute at the given time
|
---|
| 179 | node.InsertionIndex = -1;
|
---|
| 180 | ScheduleQ.OnNodeUpdated(node);
|
---|
| 181 | return Run(stopEvent);
|
---|
| 182 | }
|
---|
| 183 |
|
---|
| 184 | protected CancellationTokenSource _stop = null;
|
---|
| 185 | /// <summary>
|
---|
| 186 | /// Run until a certain event is processed.
|
---|
| 187 | /// </summary>
|
---|
| 188 | /// <remarks>
|
---|
| 189 | /// This simulation environment is not thread-safe, thus triggering this event outside the environment
|
---|
| 190 | /// leads to potential race conditions. Please use the <see cref="ThreadSafeSimulation"/> environment in case you
|
---|
| 191 | /// require this functionality. Note that the performance of <see cref="ThreadSafeSimulation"/> is lower due to locking.
|
---|
| 192 | ///
|
---|
| 193 | /// For real-time based termination, you can also call <see cref="StopAsync"/> which sets a flag indicating the simulation
|
---|
| 194 | /// to stop before processing the next event.
|
---|
| 195 | /// </remarks>
|
---|
| 196 | /// <param name="stopEvent">The event that stops the simulation.</param>
|
---|
| 197 | /// <returns></returns>
|
---|
| 198 | public virtual object Run(Event stopEvent = null) {
|
---|
| 199 | _stop = new CancellationTokenSource();
|
---|
| 200 | if (stopEvent != null) {
|
---|
| 201 | if (stopEvent.IsProcessed) {
|
---|
| 202 | return stopEvent.Value;
|
---|
| 203 | }
|
---|
| 204 | stopEvent.AddCallback(StopSimulation);
|
---|
| 205 | }
|
---|
| 206 | OnRunStarted();
|
---|
| 207 | try {
|
---|
| 208 | var stop = ScheduleQ.Count == 0 || _stop.IsCancellationRequested;
|
---|
| 209 | while (!stop) {
|
---|
| 210 | Step();
|
---|
| 211 | stop = ScheduleQ.Count == 0 || _stop.IsCancellationRequested;
|
---|
| 212 | }
|
---|
| 213 | } catch (StopSimulationException e) { OnRunFinished(); return e.Value; }
|
---|
| 214 | OnRunFinished();
|
---|
| 215 | if (stopEvent == null) return null;
|
---|
| 216 | if (!_stop.IsCancellationRequested && !stopEvent.IsTriggered) throw new InvalidOperationException("No scheduled events left but \"until\" event was not triggered.");
|
---|
| 217 | return stopEvent.Value;
|
---|
| 218 | }
|
---|
| 219 |
|
---|
| 220 | public virtual void StopAsync() {
|
---|
| 221 | _stop?.Cancel();
|
---|
| 222 | }
|
---|
| 223 |
|
---|
| 224 | public event EventHandler RunStarted;
|
---|
| 225 | protected void OnRunStarted() {
|
---|
| 226 | RunStarted?.Invoke(this, EventArgs.Empty);
|
---|
| 227 | }
|
---|
| 228 |
|
---|
| 229 | public event EventHandler RunFinished;
|
---|
| 230 | protected void OnRunFinished() {
|
---|
| 231 | RunFinished?.Invoke(this, EventArgs.Empty);
|
---|
| 232 | }
|
---|
| 233 |
|
---|
| 234 | /// <summary>
|
---|
| 235 | /// Performs a single step of the simulation, i.e. process a single event
|
---|
| 236 | /// </summary>
|
---|
| 237 | /// <remarks>
|
---|
| 238 | /// This method is not thread-safe
|
---|
| 239 | /// </remarks>
|
---|
| 240 | public virtual void Step() {
|
---|
| 241 | Event evt;
|
---|
| 242 | var next = ScheduleQ.Dequeue();
|
---|
| 243 | Now = next.PrimaryPriority;
|
---|
| 244 | evt = next.Event;
|
---|
| 245 | evt.Process();
|
---|
| 246 | ProcessedEvents++;
|
---|
| 247 | }
|
---|
| 248 |
|
---|
| 249 | /// <summary>
|
---|
| 250 | /// Peeks at the time of the next event in terms of the defined step
|
---|
| 251 | /// </summary>
|
---|
| 252 | /// <remarks>
|
---|
| 253 | /// This method is not thread-safe
|
---|
| 254 | /// </remarks>
|
---|
| 255 | public virtual double PeekD() {
|
---|
| 256 | if (ScheduleQ.Count == 0) return double.MaxValue;
|
---|
| 257 | return (Peek() - StartDate).TotalSeconds / DefaultTimeStepSeconds;
|
---|
| 258 | }
|
---|
| 259 |
|
---|
| 260 | /// <summary>
|
---|
| 261 | /// Peeks at the time of the next event
|
---|
| 262 | /// </summary>
|
---|
| 263 | /// <remarks>
|
---|
| 264 | /// This method is not thread-safe
|
---|
| 265 | /// </remarks>
|
---|
| 266 | public virtual DateTime Peek() {
|
---|
| 267 | return ScheduleQ.Count > 0 ? ScheduleQ.First.PrimaryPriority : DateTime.MaxValue;
|
---|
| 268 | }
|
---|
| 269 |
|
---|
| 270 | protected virtual void StopSimulation(Event @event) {
|
---|
| 271 | throw new StopSimulationException(@event.Value);
|
---|
| 272 | }
|
---|
| 273 |
|
---|
| 274 | public virtual void Log(string message, params object[] args) {
|
---|
| 275 | if (Logger != null)
|
---|
| 276 | Logger.WriteLine(message, args);
|
---|
| 277 | }
|
---|
| 278 |
|
---|
| 279 | #region Random number distributions
|
---|
| 280 | public double RandUniform(IRandom random, double a, double b) {
|
---|
| 281 | return a + (b - a) * random.NextDouble();
|
---|
| 282 | }
|
---|
| 283 | public double RandUniform(double a, double b) {
|
---|
| 284 | return RandUniform(Random, a, b);
|
---|
| 285 | }
|
---|
| 286 |
|
---|
| 287 | public TimeSpan RandUniform(IRandom random, TimeSpan a, TimeSpan b) {
|
---|
| 288 | return TimeSpan.FromSeconds(RandUniform(random, a.TotalSeconds, b.TotalSeconds));
|
---|
| 289 | }
|
---|
| 290 | public TimeSpan RandUniform(TimeSpan a, TimeSpan b) {
|
---|
| 291 | return RandUniform(Random, a, b);
|
---|
| 292 | }
|
---|
| 293 | public double RandTriangular(IRandom random, double low, double high) {
|
---|
| 294 | var u = random.NextDouble();
|
---|
| 295 | if (u > 0.5)
|
---|
| 296 | return high + (low - high) * Math.Sqrt(((1.0 - u) / 2));
|
---|
| 297 | return low + (high - low) * Math.Sqrt(u / 2);
|
---|
| 298 | }
|
---|
| 299 | public double RandTriangular(double low, double high) {
|
---|
| 300 | return RandTriangular(Random, low, high);
|
---|
| 301 | }
|
---|
| 302 |
|
---|
| 303 | public TimeSpan RandTriangular(IRandom random, TimeSpan low, TimeSpan high) {
|
---|
| 304 | return TimeSpan.FromSeconds(RandTriangular(random, low.TotalSeconds, high.TotalSeconds));
|
---|
| 305 | }
|
---|
| 306 | public TimeSpan RandTriangular(TimeSpan low, TimeSpan high) {
|
---|
| 307 | return RandTriangular(Random, low, high);
|
---|
| 308 | }
|
---|
| 309 |
|
---|
| 310 | public double RandTriangular(IRandom random, double low, double high, double mode) {
|
---|
| 311 | var u = random.NextDouble();
|
---|
| 312 | var c = (mode - low) / (high - low);
|
---|
| 313 | if (u > c)
|
---|
| 314 | return high + (low - high) * Math.Sqrt(((1.0 - u) * (1.0 - c)));
|
---|
| 315 | return low + (high - low) * Math.Sqrt(u * c);
|
---|
| 316 | }
|
---|
| 317 | public double RandTriangular(double low, double high, double mode) {
|
---|
| 318 | return RandTriangular(Random, low, high, mode);
|
---|
| 319 | }
|
---|
| 320 |
|
---|
| 321 | public TimeSpan RandTriangular(IRandom random, TimeSpan low, TimeSpan high, TimeSpan mode) {
|
---|
| 322 | return TimeSpan.FromSeconds(RandTriangular(random, low.TotalSeconds, high.TotalSeconds, mode.TotalSeconds));
|
---|
| 323 | }
|
---|
| 324 | public TimeSpan RandTriangular(TimeSpan low, TimeSpan high, TimeSpan mode) {
|
---|
| 325 | return RandTriangular(Random, low, high, mode);
|
---|
| 326 | }
|
---|
| 327 |
|
---|
| 328 | /// <summary>
|
---|
| 329 | /// Returns a number that is exponentially distributed given a certain mean.
|
---|
| 330 | /// </summary>
|
---|
| 331 | /// <remarks>
|
---|
| 332 | /// Unlike in other APIs here the mean should be given and not the lambda parameter.
|
---|
| 333 | /// </remarks>
|
---|
| 334 | /// <param name="random">The random number generator to use.</param>
|
---|
| 335 | /// <param name="mean">The mean(!) of the distribution is 1 / lambda.</param>
|
---|
| 336 | /// <returns>A number that is exponentially distributed</returns>
|
---|
| 337 | public double RandExponential(IRandom random, double mean) {
|
---|
| 338 | return -Math.Log(1 - random.NextDouble()) * mean;
|
---|
| 339 | }
|
---|
| 340 | /// <summary>
|
---|
| 341 | /// Returns a number that is exponentially distributed given a certain mean.
|
---|
| 342 | /// </summary>
|
---|
| 343 | /// <remarks>
|
---|
| 344 | /// Unlike in other APIs here the mean should be given and not the lambda parameter.
|
---|
| 345 | /// </remarks>
|
---|
| 346 | /// <param name="mean">The mean(!) of the distribution is 1 / lambda.</param>
|
---|
| 347 | /// <returns>A number that is exponentially distributed</returns>
|
---|
| 348 | public double RandExponential(double mean) {
|
---|
| 349 | return RandExponential(Random, mean);
|
---|
| 350 | }
|
---|
| 351 |
|
---|
| 352 | /// <summary>
|
---|
| 353 | /// Returns a timespan that is exponentially distributed given a certain mean.
|
---|
| 354 | /// </summary>
|
---|
| 355 | /// <remarks>
|
---|
| 356 | /// Unlike in other APIs here the mean should be given and not the lambda parameter.
|
---|
| 357 | /// </remarks>
|
---|
| 358 | /// <param name="random">The random number generator to use.</param>
|
---|
| 359 | /// <param name="mean">The mean(!) of the distribution is 1 / lambda.</param>
|
---|
| 360 | /// <returns>A number that is exponentially distributed</returns>
|
---|
| 361 | public TimeSpan RandExponential(IRandom random, TimeSpan mean) {
|
---|
| 362 | return TimeSpan.FromSeconds(RandExponential(random, mean.TotalSeconds));
|
---|
| 363 | }
|
---|
| 364 | /// <summary>
|
---|
| 365 | /// Returns a timespan that is exponentially distributed given a certain mean.
|
---|
| 366 | /// </summary>
|
---|
| 367 | /// <remarks>
|
---|
| 368 | /// Unlike in other APIs here the mean should be given and not the lambda parameter.
|
---|
| 369 | /// </remarks>
|
---|
| 370 | /// <param name="mean">The mean(!) of the distribution is 1 / lambda.</param>
|
---|
| 371 | /// <returns>A number that is exponentially distributed</returns>
|
---|
| 372 | public TimeSpan RandExponential(TimeSpan mean) {
|
---|
| 373 | return RandExponential(Random, mean);
|
---|
| 374 | }
|
---|
| 375 |
|
---|
| 376 | private bool useSpareNormal = false;
|
---|
| 377 | private double spareNormal = double.NaN;
|
---|
| 378 |
|
---|
| 379 | /// <summary>
|
---|
| 380 | /// Uses the Marsaglia polar method to generate a random variable
|
---|
| 381 | /// from two uniform random distributed values.
|
---|
| 382 | /// </summary>
|
---|
| 383 | /// <remarks>
|
---|
| 384 | /// Unlike <see cref="RandNormal(double, double)"/> this method does not
|
---|
| 385 | /// make use of a spare random variable. It discards the spare and thus
|
---|
| 386 | /// requires twice the number of calls to the underlying IRandom instance.
|
---|
| 387 | /// </remarks>
|
---|
| 388 | /// <param name="random">The random number generator to use.</param>
|
---|
| 389 | /// <param name="mu">The mean of the normal distribution.</param>
|
---|
| 390 | /// <param name="sigma">The standard deviation of the normal distribution.</param>
|
---|
| 391 | /// <returns>A number that is normal distributed.</returns>
|
---|
| 392 | public virtual double RandNormal(IRandom random, double mu, double sigma) {
|
---|
| 393 | return MarsagliaPolar(random, mu, sigma, out _); // do not reuse the spare normal in this case, because it could be from a different RNG
|
---|
| 394 | }
|
---|
| 395 | /// <summary>
|
---|
| 396 | /// Uses the Marsaglia polar method to generate a random variable
|
---|
| 397 | /// from two uniform random distributed values.
|
---|
| 398 | /// </summary>
|
---|
| 399 | /// <remarks>
|
---|
| 400 | /// A spare random variable is generated from the second uniformly
|
---|
| 401 | /// distributed value. Thus, the two calls to the uniform random number
|
---|
| 402 | /// generator will be made only every second call.
|
---|
| 403 | /// </remarks>
|
---|
| 404 | /// <param name="mu">The mean of the normal distribution.</param>
|
---|
| 405 | /// <param name="sigma">The standard deviation of the normal distribution.</param>
|
---|
| 406 | /// <returns>A number that is normal distributed.</returns>
|
---|
| 407 | public virtual double RandNormal(double mu, double sigma) {
|
---|
| 408 | if (useSpareNormal) {
|
---|
| 409 | useSpareNormal = false;
|
---|
| 410 | return spareNormal * sigma + mu;
|
---|
| 411 | } else {
|
---|
| 412 | useSpareNormal = true;
|
---|
| 413 | return MarsagliaPolar(Random, mu, sigma, out spareNormal);
|
---|
| 414 | }
|
---|
| 415 | }
|
---|
| 416 | private double MarsagliaPolar(IRandom random, double mu, double sigma, out double spare) {
|
---|
| 417 | double u, v, s;
|
---|
| 418 | do {
|
---|
| 419 | u = random.NextDouble() * 2 - 1;
|
---|
| 420 | v = random.NextDouble() * 2 - 1;
|
---|
| 421 | s = u * u + v * v;
|
---|
| 422 | } while (s > 1 || s == 0);
|
---|
| 423 | var mul = Math.Sqrt(-2.0 * Math.Log(s) / s);
|
---|
| 424 | spare = v * mul;
|
---|
| 425 | return mu + sigma * u * mul;
|
---|
| 426 | }
|
---|
| 427 |
|
---|
| 428 | /// <summary>
|
---|
| 429 | /// Uses the Marsaglia polar method to generate a random variable
|
---|
| 430 | /// from two uniform random distributed values.
|
---|
| 431 | /// </summary>
|
---|
| 432 | /// <remarks>
|
---|
| 433 | /// A spare random variable is generated from the second uniformly
|
---|
| 434 | /// distributed value. Thus, the two calls to the uniform random number
|
---|
| 435 | /// generator will be made only every second call.
|
---|
| 436 | /// </remarks>
|
---|
| 437 | /// <param name="random">The random number generator to use.</param>
|
---|
| 438 | /// <param name="mu">The mean of the normal distribution.</param>
|
---|
| 439 | /// <param name="sigma">The standard deviation of the normal distribution.</param>
|
---|
| 440 | /// <returns>A number that is normal distributed.</returns>
|
---|
| 441 | public TimeSpan RandNormal(IRandom random, TimeSpan mu, TimeSpan sigma) {
|
---|
| 442 | return TimeSpan.FromSeconds(RandNormal(random, mu.TotalSeconds, sigma.TotalSeconds));
|
---|
| 443 | }
|
---|
| 444 | /// <summary>
|
---|
| 445 | /// Uses the Marsaglia polar method to generate a random variable
|
---|
| 446 | /// from two uniform random distributed values.
|
---|
| 447 | /// </summary>
|
---|
| 448 | /// <remarks>
|
---|
| 449 | /// A spare random variable is generated from the second uniformly
|
---|
| 450 | /// distributed value. Thus, the two calls to the uniform random number
|
---|
| 451 | /// generator will be made only every second call.
|
---|
| 452 | /// </remarks>
|
---|
| 453 | /// <param name="mu">The mean of the normal distribution.</param>
|
---|
| 454 | /// <param name="sigma">The standard deviation of the normal distribution.</param>
|
---|
| 455 | /// <returns>A number that is normal distributed.</returns>
|
---|
| 456 | public TimeSpan RandNormal(TimeSpan mu, TimeSpan sigma) {
|
---|
| 457 | return RandNormal(Random, mu, sigma);
|
---|
| 458 | }
|
---|
| 459 |
|
---|
| 460 | public double RandNormalPositive(IRandom random, double mu, double sigma) {
|
---|
| 461 | double val;
|
---|
| 462 | do {
|
---|
| 463 | val = RandNormal(random, mu, sigma);
|
---|
| 464 | } while (val <= 0);
|
---|
| 465 | return val;
|
---|
| 466 | }
|
---|
| 467 | public double RandNormalPositive(double mu, double sigma) {
|
---|
| 468 | return RandNormalPositive(Random, mu, sigma);
|
---|
| 469 | }
|
---|
| 470 |
|
---|
| 471 | public TimeSpan RandNormalPositive(IRandom random, TimeSpan mu, TimeSpan sigma) {
|
---|
| 472 | return TimeSpan.FromSeconds(RandNormalPositive(random, mu.TotalSeconds, sigma.TotalSeconds));
|
---|
| 473 | }
|
---|
| 474 | public TimeSpan RandNormalPositive(TimeSpan mu, TimeSpan sigma) {
|
---|
| 475 | return RandNormalPositive(Random, mu, sigma);
|
---|
| 476 | }
|
---|
| 477 |
|
---|
| 478 | public double RandNormalNegative(IRandom random, double mu, double sigma) {
|
---|
| 479 | double val;
|
---|
| 480 | do {
|
---|
| 481 | val = RandNormal(random, mu, sigma);
|
---|
| 482 | } while (val >= 0);
|
---|
| 483 | return val;
|
---|
| 484 | }
|
---|
| 485 | public double RandNormalNegative(double mu, double sigma) {
|
---|
| 486 | return RandNormalNegative(Random, mu, sigma);
|
---|
| 487 | }
|
---|
| 488 |
|
---|
| 489 | public TimeSpan RandNormalNegative(IRandom random, TimeSpan mu, TimeSpan sigma) {
|
---|
| 490 | return TimeSpan.FromSeconds(RandNormalNegative(random, mu.TotalSeconds, sigma.TotalSeconds));
|
---|
| 491 | }
|
---|
| 492 | public TimeSpan RandNormalNegative(TimeSpan mu, TimeSpan sigma) {
|
---|
| 493 | return RandNormalNegative(Random, mu, sigma);
|
---|
| 494 | }
|
---|
| 495 |
|
---|
| 496 | /// <summary>
|
---|
| 497 | /// Returns values from a log-normal distribution with the mean
|
---|
| 498 | /// exp(mu + sigma^2 / 2)
|
---|
| 499 | /// and the standard deviation
|
---|
| 500 | /// sqrt([exp(sigma^2)-1] * exp(2 * mu + sigma^2))
|
---|
| 501 | /// </summary>
|
---|
| 502 | /// <param name="random">The random number generator to use.</param>
|
---|
| 503 | /// <param name="mu">The mu parameter of the log-normal distribution (not the mean).</param>
|
---|
| 504 | /// <param name="sigma">The sigma parameter of the log-normal distribution (not the standard deviation).</param>
|
---|
| 505 | /// <returns>A log-normal distributed random value.</returns>
|
---|
| 506 | public double RandLogNormal(IRandom random, double mu, double sigma) {
|
---|
| 507 | return Math.Exp(RandNormal(random, mu, sigma));
|
---|
| 508 | }
|
---|
| 509 | /// <summary>
|
---|
| 510 | /// Returns values from a log-normal distribution with the mean
|
---|
| 511 | /// exp(mu + sigma^2 / 2)
|
---|
| 512 | /// and the standard deviation
|
---|
| 513 | /// sqrt([exp(sigma^2)-1] * exp(2 * mu + sigma^2))
|
---|
| 514 | /// </summary>
|
---|
| 515 | /// <param name="mu">The mu parameter of the log-normal distribution (not the mean).</param>
|
---|
| 516 | /// <param name="sigma">The sigma parameter of the log-normal distribution (not the standard deviation).</param>
|
---|
| 517 | /// <returns>A log-normal distributed random value.</returns>
|
---|
| 518 | public double RandLogNormal(double mu, double sigma) {
|
---|
| 519 | return RandLogNormal(Random, mu, sigma);
|
---|
| 520 | }
|
---|
| 521 |
|
---|
| 522 | /// <summary>
|
---|
| 523 | /// Returns values from a log-normal distribution with
|
---|
| 524 | /// the mean <paramref name="mean"/> and standard deviation <paramref name="stdev"/>.
|
---|
| 525 | /// </summary>
|
---|
| 526 | /// <param name="random">The random number generator to use.</param>
|
---|
| 527 | /// <param name="mean">The distribution mean.</param>
|
---|
| 528 | /// <param name="stdev">The distribution standard deviation.</param>
|
---|
| 529 | /// <returns>A log-normal distributed random value.</returns>
|
---|
| 530 | public double RandLogNormal2(IRandom random, double mean, double stdev) {
|
---|
| 531 | if (stdev == 0) return mean;
|
---|
| 532 | var sigma = Math.Sqrt(Math.Log(stdev * stdev / (mean * mean) + 1));
|
---|
| 533 | var mu = Math.Log(mean) - 0.5 * sigma * sigma;
|
---|
| 534 | return Math.Exp(RandNormal(random, mu, sigma));
|
---|
| 535 | }
|
---|
| 536 | /// <summary>
|
---|
| 537 | /// Returns values from a log-normal distribution with
|
---|
| 538 | /// the mean <paramref name="mean"/> and standard deviation <paramref name="stdev"/>.
|
---|
| 539 | /// </summary>
|
---|
| 540 | /// <param name="mean">The distribution mean.</param>
|
---|
| 541 | /// <param name="stdev">The distribution standard deviation.</param>
|
---|
| 542 | /// <returns>A log-normal distributed random value.</returns>
|
---|
| 543 | public double RandLogNormal2(double mean, double stdev) {
|
---|
| 544 | return RandLogNormal2(Random, mean, stdev);
|
---|
| 545 | }
|
---|
| 546 |
|
---|
| 547 | /// <summary>
|
---|
| 548 | /// Returns a timespan value from a log-normal distribution with the mean
|
---|
| 549 | /// exp(mu + sigma^2 / 2)
|
---|
| 550 | /// and the standard deviation
|
---|
| 551 | /// sqrt([exp(sigma^2)-1] * exp(2 * mu + sigma^2))
|
---|
| 552 | /// </summary>
|
---|
| 553 | /// <param name="random">The random number generator to use.</param>
|
---|
| 554 | /// <param name="mu">The mu parameter of the log-normal distribution (not the mean).</param>
|
---|
| 555 | /// <param name="sigma">The sigma parameter of the log-normal distribution (not the standard deviation).</param>
|
---|
| 556 | /// <returns>A log-normal distributed random timespan.</returns>
|
---|
| 557 | public TimeSpan RandLogNormal(IRandom random, TimeSpan mu, TimeSpan sigma) {
|
---|
| 558 | return TimeSpan.FromSeconds(RandLogNormal(random, mu.TotalSeconds, sigma.TotalSeconds));
|
---|
| 559 | }
|
---|
| 560 | /// <summary>
|
---|
| 561 | /// Returns a timespan value from a log-normal distribution with the mean
|
---|
| 562 | /// exp(mu + sigma^2 / 2)
|
---|
| 563 | /// and the standard deviation
|
---|
| 564 | /// sqrt([exp(sigma^2)-1] * exp(2 * mu + sigma^2))
|
---|
| 565 | /// </summary>
|
---|
| 566 | /// <param name="mu">The mu parameter of the log-normal distribution (not the mean).</param>
|
---|
| 567 | /// <param name="sigma">The sigma parameter of the log-normal distribution (not the standard deviation).</param>
|
---|
| 568 | /// <returns>A log-normal distributed random timespan.</returns>
|
---|
| 569 | public TimeSpan RandLogNormal(TimeSpan mu, TimeSpan sigma) {
|
---|
| 570 | return RandLogNormal(Random, mu, sigma);
|
---|
| 571 | }
|
---|
| 572 |
|
---|
| 573 | /// <summary>
|
---|
| 574 | /// Returns a timespan value from a log-normal distribution with
|
---|
| 575 | /// the mean <paramref name="mean"/> and standard deviation <paramref name="stdev"/>.
|
---|
| 576 | /// </summary>
|
---|
| 577 | /// <param name="random">The random number generator to use.</param>
|
---|
| 578 | /// <param name="mean">The distribution mean.</param>
|
---|
| 579 | /// <param name="stdev">The distribution standard deviation.</param>
|
---|
| 580 | /// <returns>A log-normal distributed random timespan.</returns>
|
---|
| 581 | public TimeSpan RandLogNormal2(IRandom random, TimeSpan mean, TimeSpan stdev) {
|
---|
| 582 | return TimeSpan.FromSeconds(RandLogNormal2(random, mean.TotalSeconds, stdev.TotalSeconds));
|
---|
| 583 | }
|
---|
| 584 | /// <summary>
|
---|
| 585 | /// Returns a timespan value from a log-normal distribution with
|
---|
| 586 | /// the mean <paramref name="mean"/> and standard deviation <paramref name="stdev"/>.
|
---|
| 587 | /// </summary>
|
---|
| 588 | /// <param name="mean">The distribution mean.</param>
|
---|
| 589 | /// <param name="stdev">The distribution standard deviation.</param>
|
---|
| 590 | /// <returns>A log-normal distributed random timespan.</returns>
|
---|
| 591 | public TimeSpan RandLogNormal2(TimeSpan mean, TimeSpan stdev) {
|
---|
| 592 | return RandLogNormal2(Random, mean, stdev);
|
---|
| 593 | }
|
---|
| 594 |
|
---|
| 595 | public double RandCauchy(IRandom random, double x0, double gamma) {
|
---|
| 596 | return x0 + gamma * Math.Tan(Math.PI * (random.NextDouble() - 0.5));
|
---|
| 597 | }
|
---|
| 598 | public double RandCauchy(double x0, double gamma) {
|
---|
| 599 | return RandCauchy(Random, x0, gamma);
|
---|
| 600 | }
|
---|
| 601 |
|
---|
| 602 | public TimeSpan RandCauchy(IRandom random, TimeSpan x0, TimeSpan gamma) {
|
---|
| 603 | return TimeSpan.FromSeconds(RandCauchy(random, x0.TotalSeconds, gamma.TotalSeconds));
|
---|
| 604 | }
|
---|
| 605 | public TimeSpan RandCauchy(TimeSpan x0, TimeSpan gamma) {
|
---|
| 606 | return RandCauchy(Random, x0, gamma);
|
---|
| 607 | }
|
---|
| 608 |
|
---|
| 609 | public double RandWeibull(IRandom random, double alpha, double beta) {
|
---|
| 610 | return alpha * Math.Pow(-Math.Log(1 - random.NextDouble()), 1 / beta);
|
---|
| 611 | }
|
---|
| 612 | public double RandWeibull(double alpha, double beta) {
|
---|
| 613 | return RandWeibull(Random, alpha, beta);
|
---|
| 614 | }
|
---|
| 615 |
|
---|
| 616 | public TimeSpan RandWeibull(IRandom random, TimeSpan alpha, TimeSpan beta) {
|
---|
| 617 | return TimeSpan.FromSeconds(RandWeibull(random, alpha.TotalSeconds, beta.TotalSeconds));
|
---|
| 618 | }
|
---|
| 619 | public TimeSpan RandWeibull(TimeSpan alpha, TimeSpan beta) {
|
---|
| 620 | return RandWeibull(Random, alpha, beta);
|
---|
| 621 | }
|
---|
| 622 |
|
---|
| 623 |
|
---|
| 624 | /// <summary>
|
---|
| 625 | /// Generates a random sample from a given source
|
---|
| 626 | /// </summary>
|
---|
| 627 | /// <typeparam name="T">The type of the element in parameter source</typeparam>
|
---|
| 628 | /// <exception cref="ArgumentException">
|
---|
| 629 | /// Thrown when <paramref name="source"/> and <paramref name="weights"/> have different size.
|
---|
| 630 | /// or when <paramref name="weights"/> contains an invalid or negative value.
|
---|
| 631 | /// or when <paramref name="weights"/> sum equals zero or an invalid value.
|
---|
| 632 | /// </exception>
|
---|
| 633 | /// <param name="random">The random number generator to use.</param>
|
---|
| 634 | /// <param name="source">a random sample is generated from its elements.</param>
|
---|
| 635 | /// <param name="weights">The weight associated with each entry in source.</param>
|
---|
| 636 | /// <returns>The generated random samples</returns>
|
---|
| 637 | public T RandChoice<T>(IRandom random, IList<T> source, IList<double> weights) {
|
---|
| 638 | if (source.Count != weights.Count) {
|
---|
| 639 | throw new ArgumentException("source and weights must have same size");
|
---|
| 640 | }
|
---|
| 641 |
|
---|
| 642 | double totalW = 0;
|
---|
| 643 | foreach (var w in weights) {
|
---|
| 644 | if (w < 0) {
|
---|
| 645 | throw new ArgumentException("weight values must be non-negative", nameof(weights));
|
---|
| 646 | }
|
---|
| 647 | totalW += w;
|
---|
| 648 | }
|
---|
| 649 |
|
---|
| 650 | if (double.IsNaN(totalW) || double.IsInfinity(totalW))
|
---|
| 651 | throw new ArgumentException("Not a valid weight", nameof(weights));
|
---|
| 652 | if (totalW == 0)
|
---|
| 653 | throw new ArgumentException("total weight must be greater than 0", nameof(weights));
|
---|
| 654 |
|
---|
| 655 | var rnd = random.NextDouble();
|
---|
| 656 | double aggWeight = 0;
|
---|
| 657 | int idx = 0;
|
---|
| 658 | foreach (var w in weights) {
|
---|
| 659 | if (w > 0) {
|
---|
| 660 | aggWeight += (w / totalW);
|
---|
| 661 | if (rnd <= aggWeight) {
|
---|
| 662 | break;
|
---|
| 663 | }
|
---|
| 664 | }
|
---|
| 665 | idx++;
|
---|
| 666 | }
|
---|
| 667 | return source[idx];
|
---|
| 668 | }
|
---|
| 669 | /// <summary>
|
---|
| 670 | /// Generates a random sample from a given source
|
---|
| 671 | /// </summary>
|
---|
| 672 | /// <typeparam name="T">The type of the element in parameter source</typeparam>
|
---|
| 673 | /// <exception cref="ArgumentException">
|
---|
| 674 | /// Thrown when <paramref name="source"/> and <paramref name="weights"/> have different size.
|
---|
| 675 | /// or when <paramref name="weights"/> contains an invalid or negative value.
|
---|
| 676 | /// or when <paramref name="weights"/> sum equals zero
|
---|
| 677 | /// </exception>
|
---|
| 678 | /// <param name="source">a random sample is generated from its elements.</param>
|
---|
| 679 | /// <param name="weights">The weight associated with each entry in source.</param>
|
---|
| 680 | /// <returns>The generated random samples</returns>
|
---|
| 681 | public T RandChoice<T>(IList<T> source, IList<double> weights) {
|
---|
| 682 | return RandChoice(Random, source, weights);
|
---|
| 683 | }
|
---|
| 684 |
|
---|
| 685 | #endregion
|
---|
| 686 |
|
---|
| 687 | #region Random timeouts
|
---|
| 688 | public Timeout TimeoutUniformD(IRandom random, double a, double b) {
|
---|
| 689 | return new Timeout(this, ToTimeSpan(RandUniform(random, a, b)));
|
---|
| 690 | }
|
---|
| 691 | public Timeout TimeoutUniformD(double a, double b) {
|
---|
| 692 | return TimeoutUniformD(Random, a, b);
|
---|
| 693 | }
|
---|
| 694 |
|
---|
| 695 | public Timeout TimeoutUniform(IRandom random, TimeSpan a, TimeSpan b) {
|
---|
| 696 | return new Timeout(this, RandUniform(random, a, b));
|
---|
| 697 | }
|
---|
| 698 | public Timeout TimeoutUniform(TimeSpan a, TimeSpan b) {
|
---|
| 699 | return TimeoutUniform(Random, a, b);
|
---|
| 700 | }
|
---|
| 701 |
|
---|
| 702 | public Timeout TimeoutTriangularD(IRandom random, double low, double high) {
|
---|
| 703 | return new Timeout(this, ToTimeSpan(RandTriangular(random, low, high)));
|
---|
| 704 | }
|
---|
| 705 | public Timeout TimeoutTriangularD(double low, double high) {
|
---|
| 706 | return TimeoutTriangularD(Random, low, high);
|
---|
| 707 | }
|
---|
| 708 |
|
---|
| 709 | public Timeout TimeoutTriangular(IRandom random, TimeSpan low, TimeSpan high) {
|
---|
| 710 | return new Timeout(this, RandTriangular(random, low, high));
|
---|
| 711 | }
|
---|
| 712 | public Timeout TimeoutTriangular(TimeSpan low, TimeSpan high) {
|
---|
| 713 | return TimeoutTriangular(Random, low, high);
|
---|
| 714 | }
|
---|
| 715 |
|
---|
| 716 | public Timeout TimeoutTriangularD(IRandom random, double low, double high, double mode) {
|
---|
| 717 | return new Timeout(this, ToTimeSpan(RandTriangular(random, low, high, mode)));
|
---|
| 718 | }
|
---|
| 719 | public Timeout TimeoutTriangularD(double low, double high, double mode) {
|
---|
| 720 | return TimeoutTriangularD(Random, low, high, mode);
|
---|
| 721 | }
|
---|
| 722 |
|
---|
| 723 | public Timeout TimeoutTriangular(IRandom random, TimeSpan low, TimeSpan high, TimeSpan mode) {
|
---|
| 724 | return new Timeout(this, RandTriangular(random, low, high, mode));
|
---|
| 725 | }
|
---|
| 726 | public Timeout TimeoutTriangular(TimeSpan low, TimeSpan high, TimeSpan mode) {
|
---|
| 727 | return TimeoutTriangular(Random, low, high, mode);
|
---|
| 728 | }
|
---|
| 729 |
|
---|
| 730 | public Timeout TimeoutExponentialD(IRandom random, double mean) {
|
---|
| 731 | return new Timeout(this, ToTimeSpan(RandExponential(random, mean)));
|
---|
| 732 | }
|
---|
| 733 | public Timeout TimeoutExponentialD(double mean) {
|
---|
| 734 | return TimeoutExponentialD(Random, mean);
|
---|
| 735 | }
|
---|
| 736 |
|
---|
| 737 | public Timeout TimeoutExponential(IRandom random, TimeSpan mean) {
|
---|
| 738 | return new Timeout(this, RandExponential(random, mean));
|
---|
| 739 | }
|
---|
| 740 | public Timeout TimeoutExponential(TimeSpan mean) {
|
---|
| 741 | return TimeoutExponential(Random, mean);
|
---|
| 742 | }
|
---|
| 743 |
|
---|
| 744 | public Timeout TimeoutNormalPositiveD(IRandom random, double mu, double sigma) {
|
---|
| 745 | return new Timeout(this, ToTimeSpan(RandNormalPositive(random, mu, sigma)));
|
---|
| 746 | }
|
---|
| 747 | public Timeout TimeoutNormalPositiveD(double mu, double sigma) {
|
---|
| 748 | return TimeoutNormalPositiveD(Random, mu, sigma);
|
---|
| 749 | }
|
---|
| 750 |
|
---|
| 751 | public Timeout TimeoutNormalPositive(IRandom random, TimeSpan mu, TimeSpan sigma) {
|
---|
| 752 | return new Timeout(this, RandNormalPositive(random, mu, sigma));
|
---|
| 753 | }
|
---|
| 754 | public Timeout TimeoutNormalPositive(TimeSpan mu, TimeSpan sigma) {
|
---|
| 755 | return TimeoutNormalPositive(Random, mu, sigma);
|
---|
| 756 | }
|
---|
| 757 |
|
---|
| 758 | public Timeout TimeoutLogNormalD(IRandom random, double mu, double sigma) {
|
---|
| 759 | return new Timeout(this, ToTimeSpan(RandLogNormal(random, mu, sigma)));
|
---|
| 760 | }
|
---|
| 761 | public Timeout TimeoutLogNormalD(double mu, double sigma) {
|
---|
| 762 | return TimeoutLogNormalD(Random, mu, sigma);
|
---|
| 763 | }
|
---|
| 764 |
|
---|
| 765 | public Timeout TimeoutLogNormal2D(IRandom random, double mean, double stdev) {
|
---|
| 766 | return new Timeout(this, ToTimeSpan(RandLogNormal2(random, mean, stdev)));
|
---|
| 767 | }
|
---|
| 768 | public Timeout TimeoutLogNormal2D(double mean, double stdev) {
|
---|
| 769 | return TimeoutLogNormal2D(Random, mean, stdev);
|
---|
| 770 | }
|
---|
| 771 |
|
---|
| 772 | public Timeout TimeoutLogNormal(IRandom random, TimeSpan mu, TimeSpan sigma) {
|
---|
| 773 | return new Timeout(this, RandLogNormal(random, mu, sigma));
|
---|
| 774 | }
|
---|
| 775 | public Timeout TimeoutLogNormal(TimeSpan mu, TimeSpan sigma) {
|
---|
| 776 | return TimeoutLogNormal(Random, mu, sigma);
|
---|
| 777 | }
|
---|
| 778 |
|
---|
| 779 | public Timeout TimeoutLogNormal2(IRandom random, TimeSpan mean, TimeSpan stdev) {
|
---|
| 780 | return new Timeout(this, RandLogNormal2(random, mean, stdev));
|
---|
| 781 | }
|
---|
| 782 | public Timeout TimeoutLogNormal2(TimeSpan mean, TimeSpan stdev) {
|
---|
| 783 | return TimeoutLogNormal2(Random, mean, stdev);
|
---|
| 784 | }
|
---|
| 785 | #endregion
|
---|
| 786 | }
|
---|
| 787 |
|
---|
| 788 | /// <summary>
|
---|
| 789 | /// Provides a simulation environment that is thread-safe against manipulations of the event queue.
|
---|
| 790 | /// Its performance is somewhat lower than the non-thread-safe environment (cf. <see cref="Simulation"/>)
|
---|
| 791 | /// due to the locking involved.
|
---|
| 792 | /// </summary>
|
---|
| 793 | /// <remarks>
|
---|
| 794 | /// Please carefully consider if you must really schedule the stop event in a separate thread. You can also
|
---|
| 795 | /// call <see cref="Simulation.StopAsync"/> to request termination after the current event has been processed.
|
---|
| 796 | ///
|
---|
| 797 | /// The simulation will still run in only one thread and execute all events sequentially.
|
---|
| 798 | /// </remarks>
|
---|
| 799 | public class ThreadSafeSimulation : Simulation {
|
---|
| 800 | protected object _locker;
|
---|
| 801 |
|
---|
| 802 | public ThreadSafeSimulation() : this(new DateTime(1970, 1, 1)) { }
|
---|
| 803 | public ThreadSafeSimulation(TimeSpan? defaultStep) : this(new DateTime(1970, 1, 1), defaultStep) { }
|
---|
| 804 | public ThreadSafeSimulation(DateTime initialDateTime, TimeSpan? defaultStep = null) : this(new PcgRandom(), initialDateTime, defaultStep) { }
|
---|
| 805 | public ThreadSafeSimulation(int randomSeed, TimeSpan? defaultStep = null) : this(new DateTime(1970, 1, 1), randomSeed, defaultStep) { }
|
---|
| 806 | public ThreadSafeSimulation(DateTime initialDateTime, int randomSeed, TimeSpan? defaultStep = null) : this(new PcgRandom(randomSeed), initialDateTime, defaultStep) { }
|
---|
| 807 | public ThreadSafeSimulation(IRandom random, DateTime initialDateTime, TimeSpan? defaultStep = null) : base(random, initialDateTime, defaultStep) {
|
---|
| 808 | _locker = new object();
|
---|
| 809 | }
|
---|
| 810 |
|
---|
| 811 |
|
---|
| 812 | /// <summary>
|
---|
| 813 | /// Schedules an event to occur at the same simulation time as the call was made.
|
---|
| 814 | /// </summary>
|
---|
| 815 | /// <remarks>
|
---|
| 816 | /// This method is thread-safe against manipulations of the event queue
|
---|
| 817 | /// </remarks>
|
---|
| 818 | /// <param name="event">The event that should be scheduled.</param>
|
---|
| 819 | /// <param name="priority">The priority to rank events at the same time (smaller value = higher priority).</param>
|
---|
| 820 | public override void Schedule(Event @event, int priority = 0) {
|
---|
| 821 | lock (_locker) {
|
---|
| 822 | DoSchedule(Now, @event, priority);
|
---|
| 823 | }
|
---|
| 824 | }
|
---|
| 825 |
|
---|
| 826 | /// <summary>
|
---|
| 827 | /// Schedules an event to occur after a certain (positive) delay.
|
---|
| 828 | /// </summary>
|
---|
| 829 | /// <remarks>
|
---|
| 830 | /// This method is thread-safe against manipulations of the event queue
|
---|
| 831 | /// </remarks>
|
---|
| 832 | /// <exception cref="ArgumentException">
|
---|
| 833 | /// Thrown when <paramref name="delay"/> is negative.
|
---|
| 834 | /// </exception>
|
---|
| 835 | /// <param name="delay">The (positive) delay after which the event should be fired.</param>
|
---|
| 836 | /// <param name="event">The event that should be scheduled.</param>
|
---|
| 837 | /// <param name="priority">The priority to rank events at the same time (smaller value = higher priority).</param>
|
---|
| 838 | public override void Schedule(TimeSpan delay, Event @event, int priority = 0) {
|
---|
| 839 | if (delay < TimeSpan.Zero)
|
---|
| 840 | throw new ArgumentException("Negative delays are not allowed in Schedule(TimeSpan, Event).");
|
---|
| 841 | lock (_locker) {
|
---|
| 842 | var eventTime = Now + delay;
|
---|
| 843 | DoSchedule(eventTime, @event, priority);
|
---|
| 844 | }
|
---|
| 845 | }
|
---|
| 846 |
|
---|
| 847 | /// <summary>
|
---|
| 848 | /// Run until a certain event is processed.
|
---|
| 849 | /// </summary>
|
---|
| 850 | /// <remarks>
|
---|
| 851 | /// This method is thread-safe against manipulations of the event queue
|
---|
| 852 | /// </remarks>
|
---|
| 853 | /// <param name="stopEvent">The event that stops the simulation.</param>
|
---|
| 854 | /// <returns></returns>
|
---|
| 855 | public override object Run(Event stopEvent = null) {
|
---|
| 856 | _stop = new CancellationTokenSource();
|
---|
| 857 | if (stopEvent != null) {
|
---|
| 858 | if (stopEvent.IsProcessed) {
|
---|
| 859 | return stopEvent.Value;
|
---|
| 860 | }
|
---|
| 861 | stopEvent.AddCallback(StopSimulation);
|
---|
| 862 | }
|
---|
| 863 | OnRunStarted();
|
---|
| 864 | try {
|
---|
| 865 | var stop = false;
|
---|
| 866 | lock (_locker) {
|
---|
| 867 | stop = ScheduleQ.Count == 0 || _stop.IsCancellationRequested;
|
---|
| 868 | }
|
---|
| 869 | while (!stop) {
|
---|
| 870 | Step();
|
---|
| 871 | lock (_locker) {
|
---|
| 872 | stop = ScheduleQ.Count == 0 || _stop.IsCancellationRequested;
|
---|
| 873 | }
|
---|
| 874 | }
|
---|
| 875 | } catch (StopSimulationException e) { OnRunFinished(); return e.Value; }
|
---|
| 876 | OnRunFinished();
|
---|
| 877 | if (stopEvent == null) return null;
|
---|
| 878 | if (!_stop.IsCancellationRequested && !stopEvent.IsTriggered) throw new InvalidOperationException("No scheduled events left but \"until\" event was not triggered.");
|
---|
| 879 | return stopEvent.Value;
|
---|
| 880 | }
|
---|
| 881 |
|
---|
| 882 | public Task<object> RunAsync(TimeSpan duration) {
|
---|
| 883 | return Task.Run(() => Run(duration));
|
---|
| 884 | }
|
---|
| 885 |
|
---|
| 886 | public Task<object> RunAsync(DateTime until) {
|
---|
| 887 | return Task.Run(() => Run(until));
|
---|
| 888 | }
|
---|
| 889 |
|
---|
| 890 | /// <summary>
|
---|
| 891 | /// Run until a certain event is processed, but does not block.
|
---|
| 892 | /// </summary>
|
---|
| 893 | /// <param name="stopEvent">The event that stops the simulation.</param>
|
---|
| 894 | /// <returns></returns>
|
---|
| 895 | public Task<object> RunAsync(Event stopEvent = null) {
|
---|
| 896 | return Task.Run(() => Run(stopEvent));
|
---|
| 897 | }
|
---|
| 898 |
|
---|
| 899 | /// <summary>
|
---|
| 900 | /// Performs a single step of the simulation, i.e. process a single event
|
---|
| 901 | /// </summary>
|
---|
| 902 | /// <remarks>
|
---|
| 903 | /// This method is thread-safe against manipulations of the event queue
|
---|
| 904 | /// </remarks>
|
---|
| 905 | public override void Step() {
|
---|
| 906 | Event evt;
|
---|
| 907 | lock (_locker) {
|
---|
| 908 | var next = ScheduleQ.Dequeue();
|
---|
| 909 | Now = next.PrimaryPriority;
|
---|
| 910 | evt = next.Event;
|
---|
| 911 | }
|
---|
| 912 | evt.Process();
|
---|
| 913 | ProcessedEvents++;
|
---|
| 914 | }
|
---|
| 915 |
|
---|
| 916 | /// <summary>
|
---|
| 917 | /// Peeks at the time of the next event in terms of the defined step
|
---|
| 918 | /// </summary>
|
---|
| 919 | /// <remarks>
|
---|
| 920 | /// This method is thread-safe against manipulations of the event queue
|
---|
| 921 | /// </remarks>
|
---|
| 922 | public override double PeekD() {
|
---|
| 923 | lock (_locker) {
|
---|
| 924 | if (ScheduleQ.Count == 0) return double.MaxValue;
|
---|
| 925 | return (Peek() - StartDate).TotalSeconds / DefaultTimeStepSeconds;
|
---|
| 926 | }
|
---|
| 927 | }
|
---|
| 928 |
|
---|
| 929 | /// <summary>
|
---|
| 930 | /// Peeks at the time of the next event
|
---|
| 931 | /// </summary>
|
---|
| 932 | /// <remarks>
|
---|
| 933 | /// This method is thread-safe against manipulations of the event queue
|
---|
| 934 | /// </remarks>
|
---|
| 935 | public override DateTime Peek() {
|
---|
| 936 | lock (_locker) {
|
---|
| 937 | return ScheduleQ.Count > 0 ? ScheduleQ.First.PrimaryPriority : DateTime.MaxValue;
|
---|
| 938 | }
|
---|
| 939 | }
|
---|
| 940 | }
|
---|
| 941 |
|
---|
| 942 | /// <summary>
|
---|
| 943 | /// Provides a simulation environment where delays in simulation time may result in a similar
|
---|
| 944 | /// delay in wall-clock time. The environment is not an actual realtime simulation environment
|
---|
| 945 | /// in that there is no guarantee that 3 seconds in model time are also exactly 3 seconds in
|
---|
| 946 | /// observed wall-clock time. This simulation environment is a bit slower, as the overhead of
|
---|
| 947 | /// the simulation kernel (event creation, queuing, processing, etc.) is not accounted for.
|
---|
| 948 | ///
|
---|
| 949 | /// However, it features a switch between virtual and realtime, thus allowing it to be used
|
---|
| 950 | /// in contexts where realtime is only necessary sometimes (e.g. during interaction with
|
---|
| 951 | /// long-running co-processes). Such use cases may arise in simulation control problems.
|
---|
| 952 | /// </summary>
|
---|
| 953 | public class PseudoRealtimeSimulation : ThreadSafeSimulation {
|
---|
| 954 | public const double DefaultRealtimeScale = 1;
|
---|
| 955 |
|
---|
| 956 | /// <summary>
|
---|
| 957 | /// The scale at which the simulation runs in comparison to realtime. A value smaller
|
---|
| 958 | /// than 1 results in longer-than-realtime delays, while a value larger than 1 results
|
---|
| 959 | /// in shorter-than-realtime delays. A value of exactly 1 is realtime.
|
---|
| 960 | /// </summary>
|
---|
| 961 | public double? RealtimeScale { get; protected set; } = DefaultRealtimeScale;
|
---|
| 962 | /// <summary>
|
---|
| 963 | /// Whether a non-null <see cref="RealtimeScale"/> has been set.
|
---|
| 964 | /// </summary>
|
---|
| 965 | public bool IsRunningInRealtime => RealtimeScale.HasValue;
|
---|
| 966 |
|
---|
| 967 | private object _timeLocker = new object();
|
---|
| 968 | /// <summary>
|
---|
| 969 | /// The current model time. Note that, while in realtime, this may continuously change.
|
---|
| 970 | /// </summary>
|
---|
| 971 | public override DateTime Now {
|
---|
| 972 | get { lock (_timeLocker) { return base.Now + _rtDelayTime.Elapsed; } }
|
---|
| 973 | protected set => base.Now = value;
|
---|
| 974 | }
|
---|
| 975 |
|
---|
| 976 | protected CancellationTokenSource _rtDelayCtrl = null;
|
---|
| 977 | protected Stopwatch _rtDelayTime = new Stopwatch();
|
---|
| 978 |
|
---|
| 979 |
|
---|
| 980 | public PseudoRealtimeSimulation() : this(new DateTime(1970, 1, 1)) { }
|
---|
| 981 | public PseudoRealtimeSimulation(TimeSpan? defaultStep) : this(new DateTime(1970, 1, 1), defaultStep) { }
|
---|
| 982 | public PseudoRealtimeSimulation(DateTime initialDateTime, TimeSpan? defaultStep = null) : this(new PcgRandom(), initialDateTime, defaultStep) { }
|
---|
| 983 | public PseudoRealtimeSimulation(int randomSeed, TimeSpan? defaultStep = null) : this(new DateTime(1970, 1, 1), randomSeed, defaultStep) { }
|
---|
| 984 | public PseudoRealtimeSimulation(DateTime initialDateTime, int randomSeed, TimeSpan? defaultStep = null) : this(new PcgRandom(randomSeed), initialDateTime, defaultStep) { }
|
---|
| 985 | public PseudoRealtimeSimulation(IRandom random, DateTime initialDateTime, TimeSpan? defaultStep = null) : base(random, initialDateTime, defaultStep) { }
|
---|
| 986 |
|
---|
| 987 | protected override EventQueueNode DoSchedule(DateTime date, Event @event, int priority = 0) {
|
---|
| 988 | if (ScheduleQ.Count > 0 && date < ScheduleQ.First.PrimaryPriority) _rtDelayCtrl?.Cancel();
|
---|
| 989 | return base.DoSchedule(date, @event, priority);
|
---|
| 990 | }
|
---|
| 991 |
|
---|
| 992 | public override void Step() {
|
---|
| 993 | var delay = TimeSpan.Zero;
|
---|
| 994 | double? rtScale = null;
|
---|
| 995 | lock (_locker) {
|
---|
| 996 | if (IsRunningInRealtime) {
|
---|
| 997 | rtScale = RealtimeScale;
|
---|
| 998 | var next = ScheduleQ.First.PrimaryPriority;
|
---|
| 999 | delay = next - base.Now;
|
---|
| 1000 | if (rtScale.Value != 1.0) delay = TimeSpan.FromMilliseconds(delay.TotalMilliseconds / rtScale.Value);
|
---|
| 1001 | _rtDelayCtrl = CancellationTokenSource.CreateLinkedTokenSource(_stop.Token);
|
---|
| 1002 | }
|
---|
| 1003 | }
|
---|
| 1004 |
|
---|
| 1005 | if (delay > TimeSpan.Zero) {
|
---|
| 1006 | _rtDelayTime.Start();
|
---|
| 1007 | Task.Delay(delay, _rtDelayCtrl.Token).ContinueWith(_ => { }).Wait();
|
---|
| 1008 | _rtDelayTime.Stop();
|
---|
| 1009 | var observed = _rtDelayTime.Elapsed;
|
---|
| 1010 |
|
---|
| 1011 | lock (_locker) {
|
---|
| 1012 | if (rtScale.Value != 1.0) observed = TimeSpan.FromMilliseconds(observed.TotalMilliseconds / rtScale.Value);
|
---|
| 1013 | if (_rtDelayCtrl.IsCancellationRequested && observed < delay) {
|
---|
| 1014 | lock (_timeLocker) {
|
---|
| 1015 | Now = base.Now + observed;
|
---|
| 1016 | _rtDelayTime.Reset();
|
---|
| 1017 | }
|
---|
| 1018 | return; // next event is not processed, step is not actually completed
|
---|
| 1019 | }
|
---|
| 1020 | }
|
---|
| 1021 | }
|
---|
| 1022 |
|
---|
| 1023 | Event evt;
|
---|
| 1024 | lock (_locker) {
|
---|
| 1025 | var next = ScheduleQ.Dequeue();
|
---|
| 1026 | lock (_timeLocker) {
|
---|
| 1027 | _rtDelayTime.Reset();
|
---|
| 1028 | Now = next.PrimaryPriority;
|
---|
| 1029 | }
|
---|
| 1030 | evt = next.Event;
|
---|
| 1031 | }
|
---|
| 1032 | evt.Process();
|
---|
| 1033 | ProcessedEvents++;
|
---|
| 1034 | }
|
---|
| 1035 |
|
---|
| 1036 | /// <summary>
|
---|
| 1037 | /// Switches the simulation to virtual time mode, i.e., running as fast as possible.
|
---|
| 1038 | /// In this mode, events are processed without delay just like in a <see cref="ThreadSafeSimulation"/>.
|
---|
| 1039 | /// </summary>
|
---|
| 1040 | /// <remarks>
|
---|
| 1041 | /// An ongoing real-time delay is being canceled when this method is called. Usually, this
|
---|
| 1042 | /// is only the case when this method is called from a thread other than the main simulation thread.
|
---|
| 1043 | ///
|
---|
| 1044 | /// If the simulation is already in virtual time mode, this method has no effect.
|
---|
| 1045 | /// </remarks>
|
---|
| 1046 | public virtual void SetVirtualtime() {
|
---|
| 1047 | lock (_locker) {
|
---|
| 1048 | if (!IsRunningInRealtime) return;
|
---|
| 1049 | RealtimeScale = null;
|
---|
| 1050 | _rtDelayCtrl?.Cancel();
|
---|
| 1051 | }
|
---|
| 1052 | }
|
---|
| 1053 |
|
---|
| 1054 | /// <summary>
|
---|
| 1055 | /// Switches the simulation to real time mode. The real time factor of
|
---|
| 1056 | /// this default mode is configurable.
|
---|
| 1057 | /// </summary>
|
---|
| 1058 | /// <remarks>
|
---|
| 1059 | /// If this method is called while running in real-time mode, but given a different
|
---|
| 1060 | /// <paramref name="realtimeScale"/>, the current delay is canceled and the remaining
|
---|
| 1061 | /// time is delayed using the new time factor.
|
---|
| 1062 | ///
|
---|
| 1063 | /// The default factor is 1, i.e., real time - a timeout of 5 seconds would cause
|
---|
| 1064 | /// a wall-clock delay of 5 seconds. With a factor of 2, the delay as measured by
|
---|
| 1065 | /// a wall clock would be 2.5 seconds, whereas a factor of 0.5, a wall-clock delay of
|
---|
| 1066 | /// 10 seconds would be observed.
|
---|
| 1067 | /// </remarks>
|
---|
| 1068 | /// <param name="realtimeScale">A value strictly greater than 0 used to scale real time events.</param>
|
---|
| 1069 | public virtual void SetRealtime(double realtimeScale = DefaultRealtimeScale) {
|
---|
| 1070 | lock (_locker) {
|
---|
| 1071 | if (realtimeScale <= 0.0) throw new ArgumentException("The simulation speed scaling factor must be strictly positive.", nameof(realtimeScale));
|
---|
| 1072 | if (IsRunningInRealtime && realtimeScale != RealtimeScale) _rtDelayCtrl?.Cancel();
|
---|
| 1073 | RealtimeScale = realtimeScale;
|
---|
| 1074 | }
|
---|
| 1075 | }
|
---|
| 1076 |
|
---|
| 1077 | /// <summary>
|
---|
| 1078 | /// This is only a convenience for mixed real- and virtual time simulations.
|
---|
| 1079 | /// It creates a new pseudo realtime process which will set the simulation
|
---|
| 1080 | /// to realtime every time it continues (e.g., if it has been set to virtual time).
|
---|
| 1081 | /// The process is automatically scheduled to be started at the current simulation time.
|
---|
| 1082 | /// </summary>
|
---|
| 1083 | /// <param name="generator">The generator function that represents the process.</param>
|
---|
| 1084 | /// <param name="priority">The priority to rank events at the same time (smaller value = higher priority).</param>
|
---|
| 1085 | /// <param name="realtimeScale">A value strictly greater than 0 used to scale real time events (1 = realtime).</param>
|
---|
| 1086 | /// <returns>The scheduled process that was created.</returns>
|
---|
| 1087 | public Process PseudoRealtimeProcess(IEnumerable<Event> generator, int priority = 0, double realtimeScale = DefaultRealtimeScale) {
|
---|
| 1088 | return new PseudoRealtimeProcess(this, generator, priority, realtimeScale);
|
---|
| 1089 | }
|
---|
| 1090 | }
|
---|
| 1091 |
|
---|
| 1092 | /// <summary>
|
---|
| 1093 | /// Environments hold the event queues, schedule and process events.
|
---|
| 1094 | /// </summary>
|
---|
| 1095 | [Obsolete("Use class Simulation or ThreadSafeSimulation instead. Due to name clashes with System.Environment the class SimSharp.Environment is being outphased.")]
|
---|
| 1096 | public class Environment : ThreadSafeSimulation {
|
---|
| 1097 | public Environment()
|
---|
| 1098 | : base() {
|
---|
| 1099 | Random = new SystemRandom();
|
---|
| 1100 | }
|
---|
| 1101 | public Environment(TimeSpan? defaultStep)
|
---|
| 1102 | : base(defaultStep) {
|
---|
| 1103 | Random = new SystemRandom();
|
---|
| 1104 | }
|
---|
| 1105 | public Environment(int randomSeed, TimeSpan? defaultStep = null)
|
---|
| 1106 | : base(randomSeed, defaultStep) {
|
---|
| 1107 | Random = new SystemRandom(randomSeed);
|
---|
| 1108 | }
|
---|
| 1109 | public Environment(DateTime initialDateTime, TimeSpan? defaultStep = null)
|
---|
| 1110 | : base(initialDateTime, defaultStep) {
|
---|
| 1111 | Random = new SystemRandom();
|
---|
| 1112 | }
|
---|
| 1113 | public Environment(DateTime initialDateTime, int randomSeed, TimeSpan? defaultStep = null)
|
---|
| 1114 | : base(initialDateTime, randomSeed, defaultStep) {
|
---|
| 1115 | Random = new SystemRandom(randomSeed);
|
---|
| 1116 | }
|
---|
| 1117 |
|
---|
| 1118 | protected static readonly double NormalMagicConst = 4 * Math.Exp(-0.5) / Math.Sqrt(2.0);
|
---|
| 1119 | public override double RandNormal(double mu, double sigma) {
|
---|
| 1120 | return RandNormal(Random, mu, sigma);
|
---|
| 1121 | }
|
---|
| 1122 | public override double RandNormal(IRandom random, double mu, double sigma) {
|
---|
| 1123 | double z, zz, u1, u2;
|
---|
| 1124 | do {
|
---|
| 1125 | u1 = random.NextDouble();
|
---|
| 1126 | u2 = 1 - random.NextDouble();
|
---|
| 1127 | z = NormalMagicConst * (u1 - 0.5) / u2;
|
---|
| 1128 | zz = z * z / 4.0;
|
---|
| 1129 | } while (zz > -Math.Log(u2));
|
---|
| 1130 | return mu + z * sigma;
|
---|
| 1131 | }
|
---|
| 1132 | }
|
---|
| 1133 | }
|
---|