The “Choosing an Analysis Method” post from two weeks ago discussed the three macroscopic analysis methods available in Synchro: Percentile Method, HCM 2010, and HCM 2000. In addition to these choices, you can use SimTraffic for analysis.
The intention is to use Synchro and SimTraffic as companion models. Synchro can be used to determine macro level LOS and delays, then use SimTraffic to simulate realworld conditions. Not all projects require use of SimTraffic. Synchro analysis may be sufficient for a high-level planning study, for example.
- Calibrate to match real-world conditions.
- Multiple runs should be averaged to reduce the variability in results. Recording your file with one random seed is much like watching your traffic in the street for 1 hour and saying this represents all days. It is suggested that you record your model with 5 to 10 simulation runs.
- SimTraffic defaults to a 3 minute seed time and 10 minute recording time. These values may be sufficient for animation purposes, but should be increased before generating results. A 10 minute seed time and 60 minute record time are commonly used values.
Generally, microscopic models (like SimTraffic) are more realistic, however here are some disadvantages to consider:
- Large computer time and storage requirements. This can be improved by running models on a local drive instead of on a network drive, and limiting the size of the model.
- Does not optimize signal timing. Model must be re-recorded after making changes in Synchro.
- Calibration of simulation models is time consuming.
- A single error can cause inaccurate results at all study intersections.