Citation

Gaddam HK, Rao KR. Transp. Lett. 2024; 16(1): 89-102.

Copyright

(Copyright © 2024, Maney Publishing, Publisher Informa - Taylor and Francis Group)

DOI

10.1080/19427867.2022.2160559

PMID

unavailable

Abstract

.The present study aims to calibrate a novel multi-class non-lane-based continuum traffic flow model by applying advanced global search optimization algorithms such as Hybrid Search (Combination of genetic algorithm and Nelder Mead) and Generalized Pattern Search. To represent vehicular behavior in non-lane-based mixed traffic environment, the new continuum model is deduced from two-sided lateral gap car following theory along with few empirical observations. The newness of the selected model is that it mimics complex overtaking behavior of vehicles, effect of slow-moving vehicles on traffic stream and driver's anticipation behavior. The search algorithms considered in this study have been tested and compared using the real-world data. From the analysis, it is observed that the proposed algorithms are more accurate in minimizing the cost function whereas convergence speed is also found to be better. The calibrated parameters are validated using platoon dispersion characteristics of vehicles and alternative transport policy measures.

Language: en

Keywords

calibration; Macroscopic continuum model; non-lane behavior; optimization algorithms; transport policy measures