numBlockedByVehicle

Vehicle blocked time, decomposed by what was in the way: another vehicle, an occupier, or a population.

This split is what makes the whole construct validatable. A model with no spills, no closures and no picking interference must still match observed throughput, so that time is fitted into inflated task times or a depressed velocity -- the model then matches the aggregate and is wrong about the mechanism, and will give bad advice about any change that alters the obstruction rate, which is the change a study is usually commissioned to evaluate. Three separately observable quantities turn one fitted fudge into a testable claim.

Time-weighted counts, so each one's time-average is the mean number of vehicles held up by that cause, and the three sum to numTransportersBlocked. On the space rather than on the transporters, because three responses per vehicle would multiply the response count of every fleet to report what three responses per guide path report just as well.

numBlockedByIdleVehicle splits this one further and is not a fourth term in that sum.