Optimizing cycling network design in the city of Parma

Cycling is a sustainable mode of transport that helps address urban challenges such as traffic congestion and high carbon dioxide emissions. However, to encourage citizens accustomed to using private motorized vehicles or public transportation to adopt cycling, it is essential to provide suitable infrastructure that meets expectations regarding factors such as travel distance, safety, and comfort. This paper proposes an optimization strategy for designing the most suitable bicycle network for the city of Parma, Italy, based on a set of possible interventions (e.g., building new bicycle lanes or improving the quality of existing lanes) that can be implemented by the municipality under a limited budget. Given cyclists’ preferences for road characteristics including length, safety, and practicability, the optimization strategy determines both the interventions to implement and the routes to recommend for a set of origin–destination pairs, aiming to minimize the perceived costs of travel according to these weighted preferences. For this purpose, we develop a Mixed-Integer Linear Programming (MILP) formulation and propose three solution methods: a branch-and-bound algorithm and two heuristic procedures based on knapsack problems, solved using dynamic programming. Computational experiments are conducted using both real-world data from Parma and randomly generated instances to evaluate performance and to provide useful insights for infrastructure planners.