The scarcity of land in urban areas, coupled with ever growing travel demand has led to severe and prolonged traffic congestion and delays. The typical mitigation strategies include travel demand management, capacity expansion, and efficient capacity utilization. In the urban context, the essence of capacity expansion lies in the integration of various travel modes. In this work, a bi-level model is introduced to help identify the best combination of travel lanes for auto and transit usage for an urban street network with confined space. In particular, the lower-level model solvable by a successive average assignment algorithm helps determine travel lane configurations for individual street segments with due cognizance of network connectivity, upper and lower bounds of mode-specific lane widths, balancing between intersection-related in-bound and out-bound mode-specific travel lanes, etc. The travel lane configurations iteratively generated from the lower-level model are utilized as inputs for the upper-level model executed by a branch-and bound algorithm to estimate the total travel time associated with all network users. Finally, the optimal travel-way configuration for multimodal usage with the lowest network-wide total travel time is identified. A computational experiment is conducted for model application using data on a 4-square-km subarea urban network in city of Xi’an, China that covers 25 nodes, 32 links, and 5 bus routes accommodating 5,400 and 10,200 person-trips using the subarea network in 10-min off-peak and AM peak periods, respectively. For bus route configuration, options of without and with dedicated bus-only-lanes are considered. Model execution results indicate that both auto drivers and bus riders would experience reductions in origin-destination (O-D) travel time after optimization of lane number and width configuration for the auto and bus shared network and a larger extent of travel time reductions is achieved for the no bus-only lane option.

CICTP