Traffic assignment by paired alternative segments: An Ohio case study