Analytically Derived versus Numerically Derived Urban Transit Guidelines

Urban transit guidelines yield optimal network characteristics relating to factors such as stop density, network density, service frequency, and network hierarchy. These guidelines are typically derived from a simple analytical formulation of the bilevel network design problem. This optimization problem can be solved analytically only when simplifying assumptions are made regarding the demand distribution, local constraints, and travel behavior. The primary objective of this paper is to verify the validity of the assumptions underlying analytically derived guidelines. To this end, a new numerical optimization tool is developed and applied to the design of several detailed topological transit networks (including routes and line-specific properties) for the medium-sized Dutch city of Utrecht, Netherlands. Average network characteristics are derived, and a comparison is made between the existing analytically derived guidelines, the new numerically derived optimal characteristics, and the characteristics of the present network. Both analytically derived and numerically derived guidelines recommend lower stop densities, coarser networks, and higher frequencies than those found in real life. Furthermore, the introduction of zone lines would improve network quality, while express lines are beneficial only for large demand concentrations. Although existing (typically analytically derived) guidelines are found to be in line with numerically derived guidelines, the former are less suitable because local constraints are not accounted for, and only average network characteristics are computed. Therefore, the detailed network designs that are numerically derived in this study are easier to implement.