Balancing Shared Mobility Fleet Sizes: A Simulation-Driven Evolutionary Approach

Janik Muires, Lennard Sund, Tobias Ramm, Wolfgang Ketter

Keywords: Multi-Agent Simulation, Genetic Algorithm, Shared Mobility, Sustainability

TL;DR: We use simulation and evolutionary algorithms for prescriptive analytics of shared mobility tenders that strike a balance of economical, social, and environmental cost.

Abstract:

Regulators in cities face the need to enforce limits on the number of free-floating vehicle sharing schemes and vehicles. The tried-and-tested instrument for cities are tenders, for fleet sizes of the individual vehicle types. The composition of fleet sizes is often, however, guesswork or based on anecdotal evidence rather than reliable data.
Factors that are of interest include cost of operation, social equity, and environmental sustainability. Balancing them is a complex problem, but solving it could greatly support decision makers in making informed decisions for an optimal configuration of the urban mobility system.
We use a large-scale multi-agent simulation, based on empirical data from Berlin, Germany, genetic algorithm and heuristics to generate a partial solution set and discuss its applicability and boundaries.

Submission Number: 13