Solver
A named solver configuration for benchmarking. The factory creates a fresh, member-safe solver instance per cell (a new instance on every call, bound only to the supplied evaluator, with its own solver-side streams), for whatever problem the experiment hands it — the same configuration races on every problem in the grid. The harness reads each instance's configurationProperties so the configuration that actually ran is recorded rather than assumed.
Budget agnosticism: the case's factory should configure the algorithm (schedules, population sizes, replications per evaluation) but NOT its termination — the benchmark experiment wraps the factory so every created instance receives the experiment's replication-budget stopping criterion and an iteration ceiling derived from the budget. A maximum-iteration count set by the case's factory is overridden.
Parameters
a short, unique (within an experiment) identifier for the configuration; flows into cell labels, solver names, and result records
creates a fresh solver instance per benchmark cell
an optional human-readable description of the configuration