To come up with a realistic solution to a trip planning problem, you need to know the demand down to node level and the times of demand. This data can be imported from external sources or disaggregated from a zone demand matrix, using the Generate trip requests procedure.
The procedure uses a node-zone allocation, as well as a weight at the nodes that shows how the demand is distributed across the nodes. The trip request generated contains the start and destination nodes, as well as the pick-up and drop-off nodes. The trip request defines the pick-up and arrival time window within which the passenger is to be transported.
The pick-up and drop-off nodes are connected via a shortest path search performed in the network opened for the PuT Walk-TSys. The pick-up time window starts as soon as the pick-up node is reached. It ends once the maximum wait time has elapsed. The end of the arrival time window is calculated as follows:
Maximum arrival time= earliest departure time + ideal travel time + min(max((detour factor-1)*ideal travel time, always accepted detour time), maximum detour time)
The maximum wait time is not included in the latest arrival time. The maximum wait time is only meant to limit the time a passenger has to wait for a vehicle to arrive. The stochastic disaggregation of zone demand is performed using a seed random number. So when you use the same network and the same random number, the same trip requests will be generated. Vary the seed random number to obtain a temporally and spatially different distribution of the trip requests. The trip requests generated always contain integer values. "Bucket rounding” is used to ensure that in the first rank, the total demand is shown. In the second rank, the demand per origin zone and in the third rank, the demand down to OD pair level is modeled.
