You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The Vehicle Routing Problem solver and the Route solver with stop re-ordering (TSP) option both rely on first calculating an OD cost matrix of all inputs to all other inputs and use that data to configure the stop order and then subsequently calculate the travel times after the order has been determined. Unfortunately, this underlying OD cost matrix is calculated in a time neutral environment. This poses a problem for analyses done using the transit evaluator because in the time neutral case, the evaluator ignores the transit lines and reverts essentially to walking. Consequently, the walk time is used to determine the ordering instead of the transit travel time. Thus, the results of VRP and Route TSP analysis with transit are pretty wrong.
This issue is to investigate how to get around this problem. The current idea is to pre-calculate an OD Cost Matrix using transit, either at a specific point in time or using an average travel time calculated using the Calculate Travel Time Statistics tool, and inject that pre-calculated OD Cost Matrix into VRP or Route. This can't be done with python and requires ArcObjects.
The text was updated successfully, but these errors were encountered:
Just a follow-up note: The Add GTFS to a Network Dataset tool is now deprecated as the functionality is available out of the box in ArcGIS Pro. However, the same limitation applies there, and the VRP and TSP solvers throw an error preventing you from using transit networks and getting incorrect solutions.
The Vehicle Routing Problem solver and the Route solver with stop re-ordering (TSP) option both rely on first calculating an OD cost matrix of all inputs to all other inputs and use that data to configure the stop order and then subsequently calculate the travel times after the order has been determined. Unfortunately, this underlying OD cost matrix is calculated in a time neutral environment. This poses a problem for analyses done using the transit evaluator because in the time neutral case, the evaluator ignores the transit lines and reverts essentially to walking. Consequently, the walk time is used to determine the ordering instead of the transit travel time. Thus, the results of VRP and Route TSP analysis with transit are pretty wrong.
This issue is to investigate how to get around this problem. The current idea is to pre-calculate an OD Cost Matrix using transit, either at a specific point in time or using an average travel time calculated using the Calculate Travel Time Statistics tool, and inject that pre-calculated OD Cost Matrix into VRP or Route. This can't be done with python and requires ArcObjects.
The text was updated successfully, but these errors were encountered: