A comparative assessment of the LWR-IM traffic model using linear regression

Traffic models used for predictions of traffic parameters needs to be validated. This paper presents a comparative assessment to validate the predictive ability of the Lighthill-Witham-Richards - Integrated Model (LWR-IM) traffic model in simulating average delays in urban arterials using linear regression. For this purpose, the LWR-IM, TRANSYT, CTM and HCM 2000 are applied to a test intersection. Average delays are simulated by these models based on 20 different traffic scenarios. Average delays simulated by these traffic models are analyzed using the linear regression to assess how closely fitted the delays simulated by the LWR-IM with delays simulated by TRANSYT, CTM and HCM 2000. The regression reveals high degrees of correspondence with R
²
exceeding 0.9 for linear regression of average delays from the LWR-IM with predictions from the other traffic models.