/// <summary>
/// Bayesian optimization (BO) for global black box optimization problems. BO learns a model based on the initial parameter sets and scores.
/// This model is used to sample new promising parameter candiates which are evaluated and added to the existing paramter sets.
/// This process iterates several times. The method is computational expensive so is most relevant for expensive problems,
/// where each evaluation of the function to minimize takes a long time, like hyper parameter tuning a machine learning method.
/// But in that case it can usually reduce the number of iterations required to reach a good solution compared to less sophisticated methods.
/// Implementation loosely based on:
/// http://www.cs.ubc.ca/~hutter/papers/10-TR-SMAC.pdf
/// https://papers.nips.cc/paper/4522-practical-bayesian-optimization-of-machine-learning-algorithms.pdf
/// https://papers.nips.cc/paper/4443-algorithms-for-hyper-parameter-optimization.pdf
/// </summary>
/// <param name="parameters">A list of parameter bounds for each optimization parameter</param>
/// <param name="maxIterations">Maximum number of iterations. MaxIteration * numberOfCandidatesEvaluatedPrIteration = totalFunctionEvaluations</param>
/// <param name="numberOfStartingPoints">Number of randomly created starting points to use for the initial model in the first iteration (default is 5)</param>
/// <param name="numberOfCandidatesEvaluatedPrIteration">How many candiate parameter set should by sampled from the model in each iteration.
/// The parameter sets are inlcuded in order of most promissing outcome (default is 1)</param>
/// <param name="seed">Seed for the random initialization</param>
public BayesianOptimizer(ParameterBounds[] parameters, int maxIterations, int numberOfStartingPoints = 5, int numberOfCandidatesEvaluatedPrIteration = 1, int seed = 42)
{
if (parameters == null)
{
throw new ArgumentNullException("parameters");
}
if (maxIterations <= 0)
{
throw new ArgumentNullException("maxIterations must be at least 1");
}
if (numberOfStartingPoints < 1)
{
throw new ArgumentNullException("numberOfParticles must be at least 1");
}
m_parameters = parameters;
m_maxIterations = maxIterations;
m_numberOfStartingPoints = numberOfStartingPoints;
m_numberOfCandidatesEvaluatedPrIteration = numberOfCandidatesEvaluatedPrIteration;
m_sampler = new RandomUniform(seed);
// Hyper parameters for regression extra trees learner. These are based on the values suggested in http://www.cs.ubc.ca/~hutter/papers/10-TR-SMAC.pdf.
// However, according to the author Frank Hutter, the hyper parameters for the forest model should not matter that much.
m_learner = new RegressionExtremelyRandomizedTreesLearner(30, 10, 2000, parameters.Length, 1e-6, 1.0, 42, false);
// optimizer for finding maximum expectation (most promissing hyper parameters) from extra trees model.
m_maximizer = new RandomSearchOptimizer(m_parameters, 1000, 42, false);
// acquisition function to maximize,
m_acquisitionFunc = AcquisitionFunctions.ExpectedImprovement;
}
/// <summary>
/// Bayesian optimization (BO) for global black box optimization problems. BO learns a model based on the initial parameter sets and scores.
/// This model is used to sample new promising parameter candiates which are evaluated and added to the existing paramter sets.
/// This process iterates several times. The method is computational expensive so is most relevant for expensive problems,
/// where each evaluation of the function to minimize takes a long time, like hyper parameter tuning a machine learning method.
/// But in that case it can usually reduce the number of iterations required to reach a good solution compared to less sophisticated methods.
/// Implementation loosely based on:
/// http://www.cs.ubc.ca/~hutter/papers/10-TR-SMAC.pdf
/// https://papers.nips.cc/paper/4522-practical-bayesian-optimization-of-machine-learning-algorithms.pdf
/// https://papers.nips.cc/paper/4443-algorithms-for-hyper-parameter-optimization.pdf
/// </summary>
/// <param name="parameters">A list of parameter bounds for each optimization parameter</param>
/// <param name="maxIterations">Maximum number of iterations. MaxIteration * numberOfCandidatesEvaluatedPrIteration = totalFunctionEvaluations</param>
/// <param name="previousParameterSets">Parameter sets from previous run</param>
/// <param name="previousParameterSetScores">Scores from from previous run corresponding to each parameter set</param>
/// <param name="numberOfCandidatesEvaluatedPrIteration">How many candiate parameter set should by sampled from the model in each iteration.
/// The parameter sets are inlcuded in order of most promissing outcome (default is 1)</param>
/// <param name="seed">Seed for the random initialization</param>
public BayesianOptimizer(ParameterBounds[] parameters, int maxIterations, List <double[]> previousParameterSets, List <double> previousParameterSetScores,
int numberOfCandidatesEvaluatedPrIteration = 1, int seed = 42)
{
if (parameters == null)
{
throw new ArgumentNullException("parameters");
}
if (maxIterations <= 0)
{
throw new ArgumentNullException("maxIterations must be at least 1");
}
if (previousParameterSets == null)
{
throw new ArgumentNullException("previousParameterSets");
}
if (previousParameterSetScores == null)
{
throw new ArgumentNullException("previousResults");
}
if (previousParameterSets.Count != previousParameterSetScores.Count)
{
throw new ArgumentException("previousParameterSets length: "
+ previousParameterSets.Count + " does not correspond with previousResults length: " + previousParameterSetScores.Count);
}
if (previousParameterSetScores.Count < 2 || previousParameterSets.Count < 2)
{
throw new ArgumentException("previousParameterSets length and previousResults length must be at least 2 and was: " + previousParameterSetScores.Count);
}
m_parameters = parameters;
m_maxIterations = maxIterations;
m_numberOfCandidatesEvaluatedPrIteration = numberOfCandidatesEvaluatedPrIteration;
m_random = new Random(seed);
// Use member to seed the random uniform sampler.
m_sampler = new RandomUniform(m_random.Next());
// Hyper parameters for regression extra trees learner. These are based on the values suggested in http://www.cs.ubc.ca/~hutter/papers/10-TR-SMAC.pdf.
// However, according to the author Frank Hutter, the hyper parameters for the forest model should not matter that much.
m_learner = new RegressionExtremelyRandomizedTreesLearner(trees: 30,
minimumSplitSize: 10,
maximumTreeDepth: 2000,
featuresPrSplit: parameters.Length,
minimumInformationGain: 1e-6,
subSampleRatio: 1.0,
seed: m_random.Next(), // Use member to seed the random uniform sampler.
runParallel: false);
// Optimizer for finding maximum expectation (most promissing hyper parameters) from extra trees model.
m_maximizer = new RandomSearchOptimizer(m_parameters, iterations: 1000,
seed: m_random.Next(), // Use member to seed the random uniform sampler.
runParallel: false);
// Acquisition function to maximize.
m_acquisitionFunc = AcquisitionFunctions.ExpectedImprovement;
m_previousParameterSets = previousParameterSets;
m_previousParameterSetScores = previousParameterSetScores;
}
public IEnumerable <IOptimizer> UnrollOptimizer(IOptimizer optimizer)
{
List <IOptimizer> optimizers = new List <IOptimizer>();
var batchRun = optimizer as BatchRun;
var experiment = optimizer as Experiment;
if (batchRun != null && batchRun.Optimizer != null)
{
for (int i = 0; i < batchRun.Repetitions; i++)
{
optimizers.AddRange(UnrollOptimizer(batchRun.Optimizer));
}
}
else if (experiment != null)
{
foreach (var opt in experiment.Optimizers)
{
optimizers.AddRange(UnrollOptimizer(opt));
}
}
else
{
optimizers.Add(optimizer);
}
return(optimizers);
}
public bool Load() {
if (loaded) {
printToConsole("A file has already been loaded.");
return false;
}
content = ContentManager.Load(filePath);
printToConsole("Loading completed!");
printToConsole("Content loaded: " + content.ToString());
optimizer = content as IOptimizer;
if (optimizer != null) {
numberOfRuns = NumberOfRuns(optimizer);
initialNumberOfRuns = optimizer.Runs.Count;
printToConsole(String.Format("Initial number of runs: {0}", initialNumberOfRuns));
PrintRuns();
optimizer.ExceptionOccurred += new EventHandler<EventArgs<Exception>>(Optimizer_Exception);
optimizer.ExecutionStateChanged += new EventHandler(Optimizer_ExecutionStateChanged);
optimizer.Stopped += new EventHandler(Optimizer_Stopped);
optimizer.ExecutionTimeChanged += new EventHandler(Optimizer_ExecutionTimeChanged);
optimizer.Runs.RowsChanged += new EventHandler(Optimizer_Runs_RowsChanged);
}
loaded = optimizer != null;
return loaded;
}
static Loss TrainStep(ObjectDetectionDataset.EntryBatch batch, Model model, IOptimizer optimizer, int classCount, ReadOnlySpan <int> strides, bool bench = false)
{
if (bench)
{
return(ComputeLosses(model, batch, classCount, strides));
}
var tape = new GradientTape();
Loss losses;
Tensor totalLoss;
using (tape.StartUsing()) {
losses = ComputeLosses(model, batch, classCount, strides);
totalLoss = losses.GIUO + losses.Conf + losses.Prob;
if (!tf.executing_eagerly() || !tf.logical_or(tf.is_inf(totalLoss), tf.is_nan(totalLoss)).numpy().any())
{
PythonList <Tensor> gradients = tape.gradient(totalLoss, model.trainable_variables);
optimizer.apply_gradients(gradients.Zip(
(PythonList <Variable>)model.trainable_variables, (g, v) => (g, v)));
}
else
{
Trace.TraceWarning("NaN/inf loss ignored");
}
}
return(losses);
}
public TypeResolverConfigurationKernel(IOptimizer optimizer, IConfigurationTypeResolver configResolver, IAssemblyScanningTypeResolver assemblyScanner)
{
Optimizer = optimizer;
AssemblyScanner = assemblyScanner;
ConfigurationResolver = configResolver;
assemblyScanner.ConfigurationKernel = this;
}
public int CompareTo(IOptimizer y)
{
if (Level == y.Level)
{
if (Priority > y.Priority)
{
return(1);
}
else if (Priority == y.Priority)
{
return(0);
}
else
{
return(-1);
}
}
else if (Level > y.Level)
{
return(1);
}
else
{
return(-1);
}
}
public bool Load()
{
if (loaded)
{
printToConsole("A file has already been loaded.");
return(false);
}
content = ContentManager.Load(filePath);
printToConsole("Loading completed!");
printToConsole("Content loaded: " + content.ToString());
optimizer = content as IOptimizer;
if (optimizer != null)
{
numberOfRuns = NumberOfRuns(optimizer);
initialNumberOfRuns = optimizer.Runs.Count;
printToConsole(String.Format("Initial number of runs: {0}", initialNumberOfRuns));
PrintRuns();
optimizer.ExceptionOccurred += new EventHandler <EventArgs <Exception> >(Optimizer_Exception);
optimizer.ExecutionStateChanged += new EventHandler(Optimizer_ExecutionStateChanged);
optimizer.Stopped += new EventHandler(Optimizer_Stopped);
optimizer.ExecutionTimeChanged += new EventHandler(Optimizer_ExecutionTimeChanged);
optimizer.Runs.RowsChanged += new EventHandler(Optimizer_Runs_RowsChanged);
}
loaded = optimizer != null;
return(loaded);
}
public TypeResolverConfigurationKernel(IOptimizer optimizer, IConfigurationTypeResolver configResolver, IAssemblyScanningTypeResolver assemblyScanner)
{
Optimizer = optimizer;
AssemblyScanner = assemblyScanner;
ConfigurationResolver = configResolver;
assemblyScanner.ConfigurationKernel = this;
}
public void Initialize(IOptimizer optimizer)
{
gamma = new NDarray <double>(1.0, InputShape);
beta = new NDarray <double>(0.0, InputShape);
gOpt = optimizer.Clone();
bOpt = optimizer.Clone();
}
public HLRunInfo(int id, IOptimizer optimizer, string filePath, int coresRequired, string savePath) {
Id = id;
Optimizer = optimizer;
FilePath = filePath;
CoresRequired = coresRequired;
SavePath = savePath;
}
static void UpdateLearningRate(IOptimizer optimizer, Variable step, LearningRateSchedule learningRateSchedule)
{
Tensor learningRate = learningRateSchedule.Get(step: step);
var optimizerLearningRate = optimizer.DynamicGet <Variable>("lr");
optimizerLearningRate.assign(learningRate);
}
public static double[] SolveInverse(
IForwardSolver forwardSolver,
IOptimizer optimizer,
SolutionDomainType solutionDomainType,
double[] dependentValues,
double[] standardDeviationValues,
InverseFitType inverseFitType,
object[] independentValues,
double[] lowerBounds,
double[] upperBounds)
{
//var opticalPropertyGuess = ((OpticalProperties[]) (independentValues[0])).First();
//var fitParameters = new double[4] { opticalPropertyGuess.Mua, opticalPropertyGuess.Musp, opticalPropertyGuess.G, opticalPropertyGuess.N };
var parametersToFit = GetParametersToFit(inverseFitType);
var opticalPropertyGuess = (OpticalProperties[])(independentValues[0]);
var fitParametersArray = opticalPropertyGuess.SelectMany(opgi => new[] { opgi.Mua, opgi.Musp, opgi.G, opgi.N }).ToArray();
var parametersToFitArray = Enumerable.Range(0, opticalPropertyGuess.Count()).SelectMany(_ => parametersToFit).ToArray();
Func <double[], object[], double[]> func = GetForwardReflectanceFuncForOptimization(forwardSolver, solutionDomainType);
var fit = optimizer.SolveWithConstraints(fitParametersArray, parametersToFitArray, lowerBounds, upperBounds, dependentValues.ToArray(),
standardDeviationValues.ToArray(), func, independentValues.ToArray());
return(fit);
}
public override void UpdateLayers(IOptimizer optimizer)
{
CurrentLayer.Learn(optimizer);
for (int i = 0; i < InputLayers.Count; i++)
{
InputLayers[i].UpdateLayers(optimizer);
}
}
public HLRunInfo(int id, IOptimizer optimizer, string filePath, int coresRequired, string savePath)
{
Id = id;
Optimizer = optimizer;
FilePath = filePath;
CoresRequired = coresRequired;
SavePath = savePath;
}
public MultilayerPerceptron(ICostFunction costFunction, IOptimizer optimizer, Func <double, int, List <Dense>, double> regularization, Func <int[], int[], double> metrics)
{
this.costFunction = costFunction;
this.optimizer = optimizer;
this.regularization = regularization;
this.metrics = metrics;
layers = new List <Dense>();
}
public FastNN(int[] numNeurons, int miniBatchSize, bool l2loss, float dropoutProb = 0)
{
this.l2loss = l2loss;
layers = new DenseLayer[numNeurons.Length - 1];
dropouts = new DropoutLayer[numNeurons.Length - 1];
activations = new Matrix <float> [numNeurons.Length];
singleActivations = new Matrix <float> [numNeurons.Length];
preActivations = new Matrix <float> [numNeurons.Length - 1];
deltas = new Matrix <float> [numNeurons.Length - 1];
this.miniBatchSize = miniBatchSize;
optimizer = new Adam(0.001F);
IActivationFunc activationFunc = new Relu();
IInitialization initialization = new HeNormal();
for (int i = 0; i < numNeurons.Length; i++)
{
activations[i] = DenseMatrix.Create(miniBatchSize, numNeurons[i], 0);
singleActivations[i] = DenseMatrix.Create(1, numNeurons[i], 0);
if (i == 0)
{
continue;
}
if (i == numNeurons.Length - 1)
{
activationFunc = new Linear();
}
preActivations[i - 1] = DenseMatrix.Create(miniBatchSize, numNeurons[i], 0);
layers[i - 1] = new DenseLayer(numNeurons[i - 1], numNeurons[i], activationFunc, initialization);
deltas[i - 1] = DenseMatrix.Create(miniBatchSize, numNeurons[i], 0);
if (dropoutProb > 0 && i < numNeurons.Length - 1)
{
dropouts[i - 1] = new DropoutLayer(miniBatchSize, numNeurons[i], dropoutProb);
}
}
computeSDOutput = false;
if (numNeurons.Last() == 2)
{
computeSDOutput = true;
}
}
private void RemoveOptimizer(IOptimizer optimizer)
{
DeregisterOptimizerEvents(optimizer);
Runs.RemoveRange(optimizer.Runs);
if (ExecutionState == ExecutionState.Prepared && !optimizers.Any(opt => opt.ExecutionState == ExecutionState.Prepared))
{
OnStopped();
}
}
/// <summary>
/// Creates an optimizer that applies the given optimizer until a fixed point is found.
/// </summary>
/// <param name="optimizer">The optimizer whose fixed point to find.</param>
/// <returns>An optimizer that applies the given optimizer until a fixed point is found.</returns>
public static IOptimizer FixedPoint(this IOptimizer optimizer)
{
if (optimizer == null)
{
throw new ArgumentNullException(nameof(optimizer));
}
return(new FixedPointOptimizer(optimizer, int.MaxValue, throwOnCycle: true));
}
public override void Initialize(IOptimizer <Type> optimizer = null)
{
gamma = NDArray <Type> .Ones(Inputs);
beta = NDArray <Type> .Zeros(Inputs);
gOpt = optimizer.Clone();
bOpt = optimizer.Clone();
}
public OptimizerTask(IOptimizer optimizer)
: base(optimizer) {
if (optimizer is Experiment || optimizer is BatchRun) {
this.ComputeInParallel = true;
} else {
this.ComputeInParallel = false;
}
}
public override void Initialize(IOptimizer <Type> optimizer)
{
double lim = 1.0 / Math.Sqrt(Inputs);
W = NumDN.Uniform <Type>(-lim, lim, Inputs, Outputs);
w0 = NDArray <Type> .Zeros(1, Outputs);
WOpt = optimizer.Clone();
w0Opt = optimizer.Clone();
}
public SingleCloudOptimizedCommandHandler(IOptimizer optimizer,
IMapper mapper,
IRepository <AWSCloudFormationTemplate> awsTemplate,
IRepository <AzureVMTemplate> azureTemplate)
{
_optimizer = optimizer;
_mapper = mapper;
_awsTemplate = awsTemplate;
_azureTemplate = azureTemplate;
}
public Executer(Mine mine, IOptimizer optimizer)
{
if (optimizer == null || mine == null)
{
throw new ArgumentNullException();
}
_optimizer = optimizer;
_mine = mine;
}
private void AddOptimizer(IOptimizer optimizer)
{
RegisterOptimizerEvents(optimizer);
Runs.AddRange(optimizer.Runs);
optimizer.Prepare();
if (ExecutionState == ExecutionState.Stopped && optimizer.ExecutionState == ExecutionState.Prepared)
{
OnPrepared();
}
}
public override double Evaluate(IChromosome chromosome)
{
try
{
var parameters = Config.Genes.Select(s =>
new MinMaxParameterSpec(min: (double)(s.MinDecimal ?? s.MinInt.Value), max: (double)(s.MaxDecimal ?? s.MaxInt.Value),
transform: Transform.Linear, parameterType: s.Precision > 0 ? ParameterType.Continuous : ParameterType.Discrete)
).ToArray();
IOptimizer optimizer = null;
if (Config.Fitness != null)
{
if (Config.Fitness.OptimizerTypeName == Enums.OptimizerTypeOptions.RandomSearch.ToString())
{
optimizer = new RandomSearchOptimizer(parameters, iterations: Config.Generations, seed: 42, maxDegreeOfParallelism: Config.MaxThreads);
}
else if (Config.Fitness.OptimizerTypeName == Enums.OptimizerTypeOptions.ParticleSwarm.ToString())
{
optimizer = new ParticleSwarmOptimizer(parameters, maxIterations: Config.Generations, numberOfParticles: Config.PopulationSize,
seed: 42, maxDegreeOfParallelism: Config.MaxThreads);
}
else if (Config.Fitness.OptimizerTypeName == Enums.OptimizerTypeOptions.Bayesian.ToString())
{
optimizer = new BayesianOptimizer(parameters, maxIterations: Config.Generations, numberOfStartingPoints: Config.PopulationSize, seed: 42);
}
else if (Config.Fitness.OptimizerTypeName == Enums.OptimizerTypeOptions.GlobalizedBoundedNelderMead.ToString())
{
optimizer = new GlobalizedBoundedNelderMeadOptimizer(parameters, maxRestarts: Config.Generations,
maxIterationsPrRestart: Config.PopulationSize, seed: 42, maxDegreeOfParallelism: Config.MaxThreads);
}
else if (Config.Fitness.OptimizerTypeName == Enums.OptimizerTypeOptions.Genetic.ToString())
{
throw new Exception("Genetic optimizer cannot be used with Sharpe Maximizer");
}
}
//todo:
// GridSearchOptimizer?
Func <double[], OptimizerResult> minimize = p => Minimize(p, (Chromosome)chromosome);
// run optimizer
var result = optimizer.OptimizeBest(minimize);
Best = ToChromosome(result, chromosome);
return(result.Error);
}
catch (Exception ex)
{
Program.Logger.Error(ex);
return(ErrorFitness);
}
}
public void Initialize(IOptimizer <U> optimizer)
{
wOpt = optimizer.Clone();
bOpt = optimizer.Clone();
double lim = 3.0 / Math.Sqrt(InputShape[0]);
weight = ND.Uniform(-lim, lim, InputShape[0], OutputShape[0]).Cast <U>();
biases = new NDarray <double>(1, OutputShape[0]).Cast <U>();
wTmp = new NDarray <double>(weight.Shape).Cast <U>();
}
public TrainerCPU(TrainType trainType, IOptimizer optimizer)
{
TrainOptimizer = optimizer;
RandomSeed = 12;
TrainTypeSetting = trainType;
BatchSize = batchSize;
random = new Random(RandomSeed);
L1Regularization = 0;
L2Regularization = 0;
}
public GroupFormationAlgorithm(List <Participant> _participants, IMatcher matcher, IEvaluator evaluator, IOptimizer optimizer, int groupSize)
{
_participants.ForEach(p => participants.Add(p.Clone()));
this.Evaluator = evaluator;
this.Matcher = matcher;
this.Optimizer = optimizer;
this.GroupSize = groupSize;
Init();
}
public Hypothesis(Predictor predictor, IOptimizer optimizer, double learningRate)
{
_costHistory = new History();
_weightHistory = new History();
_predictor = predictor;
_optimizer = new Optimizer();
_optimizer.SetLearningRate(learningRate);
_optimizer.SetOptimizer(optimizer);
}
public static IOptimizer BuildOptimizer(string type, System.Type returnClass, int incrementSize, long explicitInitialValue)
{
// FIXME: Disable this warning, or refactor without the deprecated version.
IOptimizer optimizer = BuildOptimizer(type, returnClass, incrementSize);
if (optimizer is IInitialValueAwareOptimizer)
{
((IInitialValueAwareOptimizer)optimizer).InjectInitialValue(explicitInitialValue);
}
return(optimizer);
}
private void optimizerTabPage_DragDrop(object sender, DragEventArgs e)
{
if (e.Effect != DragDropEffects.None)
{
IOptimizer optimizer = e.Data.GetData(HeuristicLab.Common.Constants.DragDropDataFormat) as IOptimizer;
if (e.Effect.HasFlag(DragDropEffects.Copy))
{
optimizer = (IOptimizer)optimizer.Clone();
}
Content.Optimizer = optimizer;
}
}
private void DeregisterOptimizerEvents(IOptimizer optimizer)
{
optimizer.ExceptionOccurred -= new EventHandler <EventArgs <Exception> >(optimizer_ExceptionOccurred);
optimizer.ExecutionTimeChanged -= new EventHandler(optimizer_ExecutionTimeChanged);
optimizer.Paused -= new EventHandler(optimizer_Paused);
optimizer.Prepared -= new EventHandler(optimizer_Prepared);
optimizer.Started -= new EventHandler(optimizer_Started);
optimizer.Stopped -= new EventHandler(optimizer_Stopped);
optimizer.Runs.CollectionReset -= new CollectionItemsChangedEventHandler <IRun>(optimizer_Runs_CollectionReset);
optimizer.Runs.ItemsAdded -= new CollectionItemsChangedEventHandler <IRun>(optimizer_Runs_ItemsAdded);
optimizer.Runs.ItemsRemoved -= new CollectionItemsChangedEventHandler <IRun>(optimizer_Runs_ItemsRemoved);
}
public OptimizerTask(IOptimizer optimizer)
: base(optimizer)
{
if (optimizer is Experiment || optimizer is BatchRun)
{
this.ComputeInParallel = true;
}
else
{
this.ComputeInParallel = false;
}
}
public virtual void Configure(IType type, IDictionary<string, string> parms, Dialect.Dialect dialect)
{
identifierType = type;
bool forceTableUse = PropertiesHelper.GetBoolean(ForceTableParam, parms, false);
string sequenceName = PropertiesHelper.GetString(SequenceParam, parms, DefaultSequenceName);
if (sequenceName.IndexOf('.') < 0)
{
string schemaName;
string catalogName;
parms.TryGetValue(PersistentIdGeneratorParmsNames.Schema, out schemaName);
parms.TryGetValue(PersistentIdGeneratorParmsNames.Catalog, out catalogName);
sequenceName = Table.Qualify(catalogName, schemaName, sequenceName);
}
int initialValue = PropertiesHelper.GetInt32(InitialParam, parms, DefaultInitialValue);
int incrementSize = PropertiesHelper.GetInt32(IncrementParam, parms, DefaultIncrementSize);
string valueColumnName = PropertiesHelper.GetString(ValueColumnParam, parms, DefaultValueColumnName);
string defOptStrategy = incrementSize <= 1 ? OptimizerFactory.None : OptimizerFactory.Pool;
string optimizationStrategy = PropertiesHelper.GetString(OptimizerParam, parms, defOptStrategy);
if (OptimizerFactory.None.Equals(optimizationStrategy) && incrementSize > 1)
{
log.Warn("config specified explicit optimizer of [" + OptimizerFactory.None + "], but [" + IncrementParam + "=" + incrementSize + "; honoring optimizer setting");
incrementSize = 1;
}
if (dialect.SupportsSequences && !forceTableUse)
{
if (OptimizerFactory.Pool.Equals(optimizationStrategy) && !dialect.SupportsPooledSequences)
{
// TODO : may even be better to fall back to a pooled table strategy here so that the db stored values remain consistent...
optimizationStrategy = OptimizerFactory.HiLo;
}
databaseStructure = new SequenceStructure(dialect, sequenceName, initialValue, incrementSize);
}
else
{
databaseStructure = new TableStructure(dialect, sequenceName, valueColumnName, initialValue, incrementSize);
}
optimizer = OptimizerFactory.BuildOptimizer(optimizationStrategy, identifierType.ReturnedClass, incrementSize);
databaseStructure.Prepare(optimizer);
}
private int NumberOfRuns(IOptimizer optimizer) {
var batchRun = optimizer as BatchRun;
var experiment = optimizer as Experiment;
if (batchRun != null && batchRun.Optimizer != null) {
return batchRun.Repetitions * NumberOfRuns(batchRun.Optimizer);
} else if (experiment != null) {
int runs = 0;
foreach (var opt in experiment.Optimizers) {
runs += NumberOfRuns(opt);
}
return runs;
} else { return 1; }
}
private void printPythonIndividuals(IOptimizer optimizer) {
var algo = optimizer as Algorithm;
if (algo == null) return;
var prob = algo.Problem as CFGPythonProblem;
if (prob == null) return;
Helper.printToConsole(String.Join(Environment.NewLine, prob.PythonProcess.GetIndidividuals()), String.Format("{0}({1})", runInfo.FileName, runInfo.Id));
}
public virtual void Prepare(IOptimizer optimizer)
{
applyIncrementSizeToSourceValues = optimizer.ApplyIncrementSizeToSourceValues;
}
private BatchRun(BatchRun original, Cloner cloner)
: base(original, cloner) {
executionState = original.executionState;
executionTime = original.executionTime;
runsExecutionTime = original.runsExecutionTime;
optimizer = cloner.Clone(original.optimizer);
repetitions = original.repetitions;
repetitionsCounter = original.repetitionsCounter;
runs = cloner.Clone(original.runs);
batchRunAction = original.batchRunAction;
Initialize();
}
private void AddOptimizer(IOptimizer optimizer) {
RegisterOptimizerEvents(optimizer);
Runs.AddRange(optimizer.Runs);
optimizer.Prepare();
if (ExecutionState == ExecutionState.Stopped && optimizer.ExecutionState == ExecutionState.Prepared)
OnPrepared();
}
private string GetGeneration(IOptimizer opt) {
var engineAlgorithm = opt as EngineAlgorithm;
if (engineAlgorithm == null) {
engineAlgorithm = opt.NestedOptimizers.Where(o => o is EngineAlgorithm
&& o.ExecutionState.Equals(HeuristicLab.Core.ExecutionState.Started)).FirstOrDefault() as EngineAlgorithm;
}
if (engineAlgorithm != null && engineAlgorithm.Results.ContainsKey("Generations")) {
return engineAlgorithm.Results["Generations"].ToString();
}
return "No generation info found.";
}
public ActionReplayOptimizer(CodeBuilder builder)
{
generalOptimizer = new GeneralOptimizer();
this.builder = builder;
}
private void DeregisterOptimizerEvents(IOptimizer optimizer) {
optimizer.ExceptionOccurred -= new EventHandler<EventArgs<Exception>>(optimizer_ExceptionOccurred);
optimizer.ExecutionTimeChanged -= new EventHandler(optimizer_ExecutionTimeChanged);
optimizer.Paused -= new EventHandler(optimizer_Paused);
optimizer.Prepared -= new EventHandler(optimizer_Prepared);
optimizer.Started -= new EventHandler(optimizer_Started);
optimizer.Stopped -= new EventHandler(optimizer_Stopped);
optimizer.Runs.CollectionReset -= new CollectionItemsChangedEventHandler<IRun>(optimizer_Runs_CollectionReset);
optimizer.Runs.ItemsAdded -= new CollectionItemsChangedEventHandler<IRun>(optimizer_Runs_ItemsAdded);
optimizer.Runs.ItemsRemoved -= new CollectionItemsChangedEventHandler<IRun>(optimizer_Runs_ItemsRemoved);
}
public void Prepare(IOptimizer optimizer)
{
_applyIncrementSizeToSourceValues = optimizer.ApplyIncrementSizeToSourceValues;
_requiresPooledSequenceGenerator = optimizer.RequiresPooledSequenceGenerator;
}
private void UpdateChildTreeNodes(TreeNodeCollection collection, IOptimizer optimizer) {
var batchRun = optimizer as BatchRun;
var experiment = optimizer as Experiment;
if (batchRun != null && batchRun.Optimizer != null) UpdateChildTreeNodes(collection, new List<IOptimizer>() { batchRun.Optimizer });
else if (experiment != null) UpdateChildTreeNodes(collection, experiment.Optimizers);
}
/// <summary>
/// Do we require a sequence with the ability to set initialValue and incrementSize
/// larger than 1?
/// </summary>
protected bool RequiresPooledSequence(int initialValue, int incrementSize, IOptimizer optimizer)
{
int sourceIncrementSize = optimizer.ApplyIncrementSizeToSourceValues ? incrementSize : 1;
return (initialValue > 1 || sourceIncrementSize > 1);
}
private TreeNode CreateTreeNode(IOptimizer optimizer) {
TreeNode node = new TreeNode(optimizer.ToString());
node.Tag = optimizer;
var algorithm = optimizer as IAlgorithm;
if (algorithm != null) {
foreach (TreeNode childNode in CreateAlgorithmChildNodes(algorithm))
node.Nodes.Add(childNode);
}
var batchRun = optimizer as BatchRun;
if (batchRun != null) {
node.Text += string.Format(" {0}/{1}", batchRun.RepetitionsCounter, batchRun.Repetitions);
}
List<TreeNode> nodes;
if (!treeNodeTagMapping.TryGetValue(optimizer, out nodes)) {
nodes = new List<TreeNode>();
treeNodeTagMapping.Add(optimizer, nodes);
RegisterNamedItemEvents(optimizer);
}
nodes.Add(node);
foreach (TreeNode childNode in node.Nodes) {
INamedItem namedItem = childNode.Tag as INamedItem;
if (namedItem != null) {
if (!treeNodeTagMapping.TryGetValue(namedItem, out nodes)) {
nodes = new List<TreeNode>();
treeNodeTagMapping.Add(namedItem, nodes);
RegisterNamedItemEvents(namedItem);
}
nodes.Add(childNode);
}
}
return node;
}
private void RemoveOptimizer(IOptimizer optimizer) {
DeregisterOptimizerEvents(optimizer);
Runs.RemoveRange(optimizer.Runs);
if (ExecutionState == ExecutionState.Prepared && !optimizers.Any(opt => opt.ExecutionState == ExecutionState.Prepared))
OnStopped();
}
private int GetSeed(IOptimizer opt) {
var pni = opt as IParameterizedItem;
if (pni == null) {
pni = opt.NestedOptimizers.Where(o => o is IParameterizedItem
&& o.ExecutionState.Equals(HeuristicLab.Core.ExecutionState.Started)).FirstOrDefault() as IParameterizedItem;
}
if (pni != null && pni.Parameters.ContainsKey("Seed")) {
return ((IntValue)pni.Parameters["Seed"].ActualValue).Value;
}
return -1;
}