Microtransit Dispatch 2D — Cheapest-Insertion Routing

Fleet Dispatch Controls

Fleet size 5 vans

Seats per van 3

Request arrival rate 18 / min

Simulation speed 1.0×

🖱 Click anywhere on the map to drop a pickup point, then click again to set the dropoff — the new request is inserted wherever it adds the least driving distance.

Live Readouts

Trips served

Waiting now

Avg wait

0.0 s

Avg detour ratio

1.00×

Fleet utilization

Total van-km

0.0

Shuttle van (colored per unit)

Pending pickup

Booked dropoff

Your click-placed pickup (armed)

How it works

This is the dynamic dial-a-ride problem (DARP): rider requests (pickup P, dropoff D) arrive continuously and must be assigned to one of several capacity-limited shuttles in real time, without knowing future requests in advance.

Every new request is inserted using the cheapest-insertion heuristic. For each vehicle v with an existing ordered stop list, the dispatcher tries every pair of insertion slots (i, j) for P and D, i < j, subject to P coming before D and the running onboard load never exceeding capacity:

cost(v,i,j) = length(route_v with P inserted at i, D inserted at j)
              − length(route_v)

assign request to  argmin over all (v,i,j) of cost(v,i,j)

The request joins whichever vehicle and slot adds the least extra driving distance — this is the same greedy strategy production microtransit dispatchers (Via, on-demand transit pilots) use as a fast approximation, since the exact DARP is NP-hard.

Fleet size / seats per van — more capacity means fewer, cheaper insertions but slower response per rider once vans fill up.
Arrival rate — push it high enough and you can watch average wait time and detour ratio climb as the greedy heuristic runs out of slack.
Detour ratio — a rider's actual ride distance divided by the direct pickup→dropoff distance; 1.00× is a private-car trip, higher values are the cost of pooling.
Click the map — drop your own pickup and dropoff and watch which van (and where in its route) the greedy insertion picks.

5 vans · cap 3 0 pending requests

⚠ Couldn't load the engineSomething went wrong initializing the canvas. Reload the page.