Traffic assignment by paired alternative segments: An Ohio case study

This paper examines the performance of a new traffic assignment algorithm, TAPAS, in a case study drawn from transportation planning practice. By identifying and adjusting route and class link flows on pairs of alternative segments, TAPAS is able to converge quickly to the user equilibrium (UE) while finding unique UE route and class link flows, which satisfy a condition of proportionality. The results show that TAPAS achieves more precise solutions compared to a commercial traffic assignment tool based on Frank-Wolfe algorithm, and it generates them with much less computational effort. The case study reveals the project benefit estimated with the TAPAS solution is about 4% lower than that given by the commercial software, which may be large enough to influence decision making. This study provides initial evidence that higher levels of precision, and unique class link flows, may be necessary for certain analyses.