Currently the speeds of isochrone agent and network analysis agent are dependent on the size of the road network data. Unnecessary slow down can occur if the road network data is much greater than the area of interest. This problem is more significant for isochrone agent as it segmentises the road network data for higher resolution.
There should be functionalities to filter the road network data to adequate range such that isochrone calculation and network analysis are efficient while remaining "correct".
Early stage ideas include:
user may specify a bounding box in the request
user may specify a maximum distance/time in the request, the agent then filter road network data accordingly
agent perform calculations iteratively with increasing amount of road network data until the result "converges"
Caveats:
For isochrone agent, a copy of the road network data is created after segmentation, and this copy is re-used in calculations. New solutions should consider re-using this table and avoid re-calculation unless there is outweighing benefit.
Currently the speeds of isochrone agent and network analysis agent are dependent on the size of the road network data. Unnecessary slow down can occur if the road network data is much greater than the area of interest. This problem is more significant for isochrone agent as it segmentises the road network data for higher resolution.
There should be functionalities to filter the road network data to adequate range such that isochrone calculation and network analysis are efficient while remaining "correct".
Early stage ideas include:
Caveats: