Class markov::LinearSweepModel

Base class for implementations which wish to have full control over how an entire sweep is performed. Implementations should.

  1. Defer to accept(delta_cost) -> bool when deciding whether to apply a proposed change (this is typically the metropolis acceptance rate) 2) Call check_lowest() whenever a potentially new lowest cost is found (check_lowest, as opposed to save_lowest, does store the state as new best only if its cost is lower than the current best) 3) Keep "cost" updated to the one corresponding to the current state (at least before every call to check_lowest() and after the sweep is completed).
Inheritance
markov::Model
markov::LinearSweepModel
Inherited Members

Methods

make_linear_sweep()

Declaration
virtual void markov::LinearSweepModel<State>::make_linear_sweep(double&cost, State&state, std::function<bool(double)>accept, std::function<void(void)>check_lowest) const =0