For calibration of Vissim Model with Field volume 1. Which results in vissim are to be used for calibration , there are data collection points, vehicle network performance evaluation results, Delay Results, Vehicle Travel Time Results? 2. And why are Vehicle(All) results vary different in other results like Delay Results, Node Results etc.
Hi J Sai, 1) you calibrate your model using data collection point and node results usually. Network performance is more an indicator for comparison between scenarios. Travel time is usually for validation once your model is calibrated. 2) Because Vissim uses a random generator. Speed, vehicle input (usually) and other settings/inputs are given in distribution and not as exact figures. The random generator basically selects within that distribution a number in each run when you use different 'random seed'. See this video for reference, I think in this one I talk about this: ua-cam.com/video/isxzbP3cSLo/v-deo.html This ensures that when you run your model multiple times you can create more reliable results and avoid having 'one off' situations. Basically you get a normal daily variation included in your model runs when you do a multirun and average the results.
thank you for sharing
Thanks for watching!
Nice video Mr. Matyas but the sound so low. Thank you so much
hmm, sorry about that, maybe something was off with my mic, did not notice it
For calibration of Vissim Model with Field volume
1. Which results in vissim are to be used for calibration , there are data collection points, vehicle network performance evaluation results, Delay Results, Vehicle Travel Time Results?
2. And why are Vehicle(All) results vary different in other results like Delay Results, Node Results etc.
Hi J Sai,
1) you calibrate your model using data collection point and node results usually. Network performance is more an indicator for comparison between scenarios. Travel time is usually for validation once your model is calibrated.
2) Because Vissim uses a random generator. Speed, vehicle input (usually) and other settings/inputs are given in distribution and not as exact figures. The random generator basically selects within that distribution a number in each run when you use different 'random seed'. See this video for reference, I think in this one I talk about this:
ua-cam.com/video/isxzbP3cSLo/v-deo.html
This ensures that when you run your model multiple times you can create more reliable results and avoid having 'one off' situations. Basically you get a normal daily variation included in your model runs when you do a multirun and average the results.