Parts | Days | Selection | Search | Updates | Downloads | Help
SOE: Fachverband Physik sozio-ökonomischer Systeme
SOE 11: Physics of Collective Mobility (joint session SOE / DY / BP / jDPG, accompanying the symposium)
SOE 11.3: Talk
Tuesday, March 21, 2017, 15:15–15:30, GÖR 226
The Simulation Unification: Towards a Python Toolbox to Model and Analyze Collective Mobility Systems — •Andreas Sorge1,2,3, Marc Timme1,2,3, and Debsankha Manik1,2 — 1MPI for Dynamics and Self-Organization, Göttingen, Germany — 2Institute for Nonlinear Dynamics, Georg-August-Universität, Göttingen, Germany — 3Organization for Research on Complex Adaptive Systems (or-cas), Göttingen, Germany
Understanding the nonlinear dynamics, scaling behavior and critical transitions of collective mobility systems is crucial to optimize system performance and individual utility. As discrete events such as pick-ups and drop-offs govern their time evolution, it is intricate to apply standard analytical and computational methods from statistical physics and nonlinear dynamics. We develop a unifying theoretical and computational framework to efficiently model and simulate such systems. Based on a formal language, we develop a Python package that allows the simulationist to assemble modules in a toolbox-like fashion. The package ships with a number of topologies (e.g. R2, any kind of network) and different dispatching rules to bundle individual requests and assign them to vehicles. The user is free to combine them and to specify their own, as well as to feed in the spatio-temporal mobility demand pattern. An analysis module provides ample statistics to aggregate the dynamics and assess the performance of the simulated instance. We envision our package to facilitate both studying collective mobility and developing the necessary tools to do so in a reproducible manner.