Simulating adaptive, personalized, multi-modal mobility in smart cities

Abstract

Smart, multi-modal transportation concepts are a key component towards smart sustainable cities. Such systems usually involve combinations of various modes of individual mobility (private cars, bicycles, walking), public transportation, and shared mobility (e.g. car sharing, car pooling). In this paper, we introduce a large-scale multi-agent simulation tool for simulating adaptive, personalized, multi-modal mobility. It is calibrated using various sources of real-world data and can be quickly adapted to new scenarios. The tool is highly modular and flexible and can be used to examine a variety of questions ranging from collective adaptation over collaborative learning to emergence and emergent behaviour. We present the design concept and architecture, showcase the adaptation to a real scenario (the city of Trento, Italy) and demonstrate an example of collaborative learning.