IBM Support

Traffic Prediction is inhibited when the road network graph is incorrectly linked

Troubleshooting


Problem

Traffic predictions are inaccurate when connections between road links are missing from the road network graph.

Symptom

Traffic predictions for one link might not reflect changes in traffic data from a geographically connected link if subgraphs are disconnected.

Cause

One source of error in Traffic Prediction (TP) is disconnected subgraphs within the link network in the shape file. TP does not consider links from disconnected subgraphs in its relationship model. If these links represent roads that are geographically connected, data from one link cannot be used in predictions for the other links.
Potential reasons for disconnected subgraphs are:

  • Variations in the end node definitions for links can result in the end node of one link not matching the start node of the next link.
  • A split in the road might be defined such that the start node of the splitting link is within the middle of the parent link, not at the end node.
The administrator of IBM Intelligent Operations for Transportation with Traffic Prediction must identify and address these issues.

Diagnosing The Problem

When subgraphs are disconnected, TP does not consider links from different disconnected subgraphs in its relationship model even if the links are physically connected. This leads to issues wherein traffic predictions for one link might not reflect changes in traffic data from a geographically connected link, or predictions might not be produced for links with disconnected neighbors.

Resolving The Problem

You can use the trans_tptool utility on the analytics server to find disconnected subgraphs. Run this utility after installation, or after updates to the link network, to analyze the information that is stored for your network and to generate a connectivity report.

For details of the utility, see the related information "FAQ: TP inhibited if road network graph incorrect".

[{"Product":{"code":"SSTMV4","label":"IBM Intelligent Transportation"},"Business Unit":{"code":"BU055","label":"Cognitive Applications"},"Component":"IBM Intelligent Transit Analytics","Platform":[{"code":"PF016","label":"Linux"}],"Version":"1.6.1","Edition":"","Line of Business":{"code":"LOB59","label":"Sustainability Software"}}]

Document Information

Modified date:
17 June 2018

UID

swg21958965