Iteration
constructor(iteration: Int, cumulativeReplications: Int, bestPenalizedObjective: Double, solverState: Map<String, Double> = emptyMap())(source)
Parameters
iteration
the solver's iteration counter at the snapshot
cumulative
the solver's cumulative requested replications at the snapshot — the x-axis for budget-normalized convergence plots
best
the penalized objective (minimization-oriented) of the best solution found so far
solver
the solver's algorithm-specific state at the snapshot, keyed by the names the solver chose — a swarm's diameter, a population's diversity, a reference distribution's coefficient of variation. These are the direct measurements of premature convergence, and the solvers emit them every iteration whether or not anyone records them. Empty unless the experiment captured solver state, and for solvers that publish none.