Stochastic dynamics of modern demand-driven mobility

Philip Keith Marszal

AutorIn

Philip Keith Marszal

Titel

Stochastic dynamics of modern demand-driven mobility

Zitierfähige Url:

https://nbn-resolving.org/urn:nbn:de:bsz:14-qucosa2-973748

Erstveröffentlichung

2025

Datum der Einreichung

21.01.2025

Datum der Verteidigung

23.05.2025

Abstract (EN)

Systems of many interacting entities often exhibit emergent properties and collective dynamics that are qualitatively distinct from those of non-interacting systems. Predicting such collective states is especially challenging if interactions are nonlinear and multifaceted. Modern mobility systems provide a compelling example of such complexity, where individuals interact both physically, for instance, on roads and at charging stations, and digitally, for instance, through route planning and mobility platforms. How do such new modes of interactions drive collective dynamics and influence the modern mobility landscape? This thesis investigates the emergence of collective dynamics in three key domains of modern mobility: ride-hailing, electric vehicle traffic, and ride-pooling. In ride-hailing, we observe and explain a form of hysteresis, where fluctuations in demand can lock the system into states of poor performance on long time scales. Ride-hailing platforms use dynamic pricing to mitigate this effect by balancing supply and demand. We reveal that dynamic pricing may also foster economic interactions among drivers that may induce coordinated driver behavior and in turn artificial supply shortages. In electric vehicle traffic, recharging events create new points of interaction at charging stations. We find that these interactions give rise to congestion waves, which display phase-separation characteristics, clustering stations into congested and uncongested states. In ride-pooling systems, vehicles are guided by central algorithms that mediate their interactions. We show that these algorithm-driven dynamics are characterized by macroscopic scaling laws. Microscopically, the mechanisms by which routes change give rise to multi-timescale diffusion. We demonstrate that the microscopic and the macroscopic are linked by a common scaling relation found for the diffusion coefficient of ride-pooling vehicles, hinting at a deeper geometric influence on the scaling laws. By revealing these dynamics and explaining the mechanisms underlying them, this work opens new avenues for understanding and shaping the mobility landscape, emphasizing the power of interdisciplinary approaches to uncover fundamental principles governing complex systems.

Freie Schlagwörter (DE)

Mobilität, Komplexe Systeme

Klassifikation (DDC)

530

Klassifikation (RVK)

UF 1950

GutachterIn