Bike network planning in limited urban space

TL;DR

This work tackles bike network planning under limited urban space by formulating BNAP, a multi-criteria optimization that explicitly accounts for the trade-off between bike and car networks on the same street space. It introduces a Pareto-based evaluation framework and an LP formulation that separates bike, car, and shared flows, augmented with auxiliary paths to guarantee car-network connectivity and demand-driven OD paths to reflect real travel patterns. Through real-data experiments (Zurich, Cambridge MA, Chicago) and synthetic tests, BNAP shows Pareto-front dominance over betweenness-based heuristics and achieves scalable city-wide planning via relaxations and iterative rounding. The framework provides urban planners with a flexible, data-driven tool to design cycling infrastructure while quantifying impacts on motorized transport, supporting more sustainable and resilient city planning.

Abstract

The lack of cycling infrastructure in urban environments hinders the adoption of cycling as a viable mode for commuting, despite the evident benefits of (e-)bikes as sustainable, efficient, and health-promoting transportation modes. Bike network planning is a tedious process, relying on heuristic computational methods that frequently overlook the broader implications of introducing new cycling infrastructure, in particular the necessity to repurpose car lanes. In this work, we call for optimizing the trade-off between bike and car networks, effectively pushing for Pareto optimality. This shift in perspective gives rise to a novel linear programming formulation towards optimal bike network allocation. Our experiments, conducted using both real-world and synthetic data, testify the effectiveness and superiority of this optimization approach compared to heuristic methods. In particular, the framework provides stakeholders with a range of lane reallocation scenarios, illustrating potential bike network enhancements and their implications for car infrastructure. Crucially, our approach is adaptable to various bikeability and car accessibility evaluation criteria, making our tool a highly flexible and scalable resource for urban planning. This paper presents an advanced decision-support framework that can significantly aid urban planners in making informed decisions on cycling infrastructure development.

Figures (14)

• Figure 1: Explanation of spatial relaxation, where flow is only allowed on the subgraph surrounding the shortest path. For each node on the shortest path between s and t, its $\eta$ closest nodes are included. Here closeness is defined in terms of the Euclidean distance in geographic space.
• Figure 2: Network layout of the real instances used to test our algorithm
• Figure 3: Pareto optimality of bike networks. Top: Algorithms are compared by their Pareto frontier. In five out of six instances, our linear programming approach outperforms methods based on the betweenness centrality. Bottom: Each point on the Pareto frontiers (top) corresponds to one plausible street network. Three examples in Cambridge MA are shown, where the bike networks differ dependent on the planning method and the number of allocated bike lanes. This is also reflected in the distance of shortest paths, where the existence of dedicated cycling infrastructure is rewarded in the perceived bike travel time.
• Figure 4: Comparison of the trade-off between bike and car travel times in different cities.
• Figure 5: Case study of rebuilding the whole city of Zurich. As many bike lanes as possible were allocated with our optimization approach. The dense bike network results in many one-way car lanes, nevertheless forming a functional and connected car network.