C# (CSharp) SwarmOps Namespace

Nested Namespaces

SwarmOps.Optimizers
SwarmOps.Problems

Classes

Name Description
FeasibleTrace Store feasibility-values during optimization runs and write them to a file afterwards. The number of iterations per optimization run must be known in advance.
FitnessPrint Prints parameters and fitness to Console. Useful in viewing the progress of meta-optimization. Works as a 'transparent' wrapper for the problem to be optimized. Note that feasibility is computed here.
FitnessTrace
FitnessTraceQuartiles Store fitness-values at intervals during optimization runs and write their quartiles to a file afterwards. This only supports a fixed number of optimization runs and iterations per run which must therefore be known in advance. Keep the number of intervals small because it requires much memory.
Globals Collection of globally used static variables.
LogSolutions Log best solutions found during optimization, that is, log parameters and their associated fitness. Transparently wraps around a problem-object.
Optimizer Base-class for an optimizer.
OptimizerWrapper Transparently wrap an Optimizer-object.
Problem Base-class for an optimization problem.
ProblemIndex Used for sorting optimization problems so that those hardest to optimize are tried first. This is used in the MetaFitness class where Pre-emptive Fitness Evaluation seeks to abort the meta-fitness evaluation as early as possible.
ProblemIndex.ProblemFitness Associate an optimization problem with a fitness.
ProblemWrapper Transparently wrap a problem-object.
Quartiles Compute quartiles.
Repeat Base-class for performing a number of optimization runs.
RepeatMin Performs a number of optimization runs and returns the minimum fitness found. Respects feasibility (constraint satisfaction.) This does NOT allow for Preemptive Fitness Evaluation!
RepeatSum Performs a number of optimization runs and returns the sum of the fitnesses. Ignores feasibility (constraint satisfaction.) This allows for Preemptive Fitness Evaluation.
Result Best solution found during optimization.
Statistics Wrapper for an optimizer providing statistics such as mean fitness achieved over a number of optimization runs, best results achieved, etc. Transparently supports the same methods as the the optimizer itself, but stores the optimization results so as to compute the statistics.
StatisticsAccumulator Compute statistical measures in accumulating manner, incl. mean, variance, std.deviation, min, max.
Tools
WeightedProblem Problem and weight pair for use in meta-optimization.