Random
class RandomWalkSolver(problemDefinition: ProblemDefinition, evaluator: EvaluatorIfc, maxIterations: Int, replicationsPerEvaluation: ReplicationPerEvaluationIfc, streamNum: Int = 0, streamProvider: RNStreamProviderIfc = KSLRandom.DefaultRNStreamProvider, name: String = "RandomWalk") : StochasticSolver(source)
A class that implements an unbiased Random Walk solver. This algorithm unconditionally explores the solution space by randomly generating and moving to a neighborhood solution at each iteration. It does not attempt to optimize, but rather freely wanders the landscape. It is primarily useful for landscape analysis, baseline benchmarking, or estimating hyperparameters for other algorithms (like Simulated Annealing).
Constructors
Link copied to clipboard
constructor(problemDefinition: ProblemDefinition, evaluator: EvaluatorIfc, maxIterations: Int, replicationsPerEvaluation: ReplicationPerEvaluationIfc, streamNum: Int = 0, streamProvider: RNStreamProviderIfc = KSLRandom.DefaultRNStreamProvider, name: String = "RandomWalk")