scaled_parallel_worker_cost

pyoptex.doe.cost_optimal.cost.scaled_parallel_worker_cost(transition_costs, factors, max_cost, execution_cost=1)

Create a transition cost function for a problem where multiple workers can work on the transition between two consecutive runs in parallel. The total transition cost is determined by the most-hard-to-change factor. The transition cost is determined by scaling the transition cost between start and stop with a base cost. See the parameters for more information.

Parameters

transition_costsdict(str, tuple(float, float, float, float) or float)

A dictionary mapping the factor name to the transition cost. The cost is a tuple with as first element the base cost of any positive transition (-1 to +1), as second element the base cost of any negative transition (+1 to -1), as third element the additional cost to positively scale between min (-1) and max (+1), and as third element the additional cost to negatively scale between max (+1) and min (-1). Categorical factors should have only a float indicating the base cost of any transition.

factorslist(Factor)

The factors for the design.

max_costfloat

The budget available for this cost function.

execution_costfloat

the execution cost of a run.

Returns

cost_fnfunc(Y, params)

The cost function.