Hi! I'm attempting to evaluate different service design choices for a special transport service/paratransit (for those of you who are unable to use general public transport due to a disability) in Stockholm, Sweden. The service design choices could be e.g. time window sizes, maximum travel times, etc.
The service is today the largest procurer of taxi trips in the country (and the fourth largest in the world!), with over two million taxi trips per year.
My problem is this: the real-world cost structure (as defined in contracts) is today that there is a fixed cost for each trip + a distance cost. BUT, the cost is zero when the vehicle is empty (so all trips to/from depots and between customers are "free"). I can recreate that after the solver is finished routing, by exporting the results to excel and doing some calculations to calculate the load (i.e. how many passengers are in the taxi) for each vehicle between the pick-up/drop-off-points. But I'm not certain that the solver would produce the same routes if the cost was 0 when the vehicles are empty.
Is there any way to define a cost-function that depends on the vehicle load?
Thanks for any advice!
