Physical Attribute-Driven NFD Approach for Bicycle Network Design

Urban street networks are increasingly expected to serve multiple modes simultaneously, yet tools for understanding how physical design choices affect aggregate traffic performance, translating those predictions into infrastructure investment decisions, and quantifying the equity trade-offs among those decisions remains limited. This work addresses these gaps through the estimation of an NFD functional form based on physical attributes of active mobility infrastructure which is then employed to develop network designs informed by varying equity objectives. The physical-infrastructure-informed NFD functional form links traffic performance, such as capacity, jam density, and flow dynamics, to spatial design characteristics including the among of area allocated to each mode, space for interaction between modes, and type of separation between modes. This form was calibrated using high-resolution simulation outputs from multimodal SUMO networks which allowed for differences in flow behavior across modes and designs to be captured in high density scenarios. A two-stage optimization framework was implemented to determine an optimal assignment of bike lane types across network links based on this functional form. The first stage maximizes the developed functional form to return optimal values of the spatial design characteristics. The second step makes use of a genetic algorithm (GA) to generate networks optimizing for flow, accessibility, and adherence to ethical frameworks. By evaluating candidate solutions using the estimated NFD, the algorithm selects infrastructure configurations that maximize network efficiency while also reducing congestion impacts on disadvantaged communities. This framework contributes an integration of macroscopic traffic modeling into network design for multimodal systems with additional equity-focused optimization. Presented by Dr. Meredith Raymer, Consultant, DB ECO North America