Background
The Lighthouse mode choice model currently includes two transit alternatives:
- Walk to Transit
- Drive to Transit
This is consistent with the transit skims currently available from TDM23. For simplicity, transit submodes (e.g., commuter rail, subway, light rail, bus) are not represented as separate mode choice alternatives. Instead, the TransCAD pathfinding and transit assignment procedures determine the optimal transit path and submode mix.
However, it is still possible to generate in-vehicle travel time (IVT) skims for specific transit submodes, such as:
- commuter_rail_ivt
- subway_ivt
- bus_ivt
These mode-specific IVT components could be incorporated into the transit utility expressions. This is a common approach for capturing differences in rider preferences across transit submodes by applying different utility coefficients to each IVT component. For example, travelers may perceive time spent on commuter rail differently from time spent on local bus service.
Adding mode-specific IVT terms could improve transit mode choice calibration while maintaining a simplified set of transit alternatives.
Goals
- Evaluate mode choice calibration and model performance using the current transit utility specification without submode-specific IVT terms.
- Assess whether mode-specific IVT skims improve model fit, transit mode shares, or transit market segmentation.
- Determine whether the benefits of including mode-specific IVT components justify the additional skim generation and model complexity.
Background
The Lighthouse mode choice model currently includes two transit alternatives:
This is consistent with the transit skims currently available from TDM23. For simplicity, transit submodes (e.g., commuter rail, subway, light rail, bus) are not represented as separate mode choice alternatives. Instead, the TransCAD pathfinding and transit assignment procedures determine the optimal transit path and submode mix.
However, it is still possible to generate in-vehicle travel time (IVT) skims for specific transit submodes, such as:
These mode-specific IVT components could be incorporated into the transit utility expressions. This is a common approach for capturing differences in rider preferences across transit submodes by applying different utility coefficients to each IVT component. For example, travelers may perceive time spent on commuter rail differently from time spent on local bus service.
Adding mode-specific IVT terms could improve transit mode choice calibration while maintaining a simplified set of transit alternatives.
Goals