DIGITAL LABORATORY OF AGENT-BASED HIGHWAY TRAFFIC MODEL

We describe our microscopic model of highway traffic and its implementation as digital laboratory. We discuss the range of experiments that can be conducted using this laboratory. Software implementation details are discussed together with the model because the implementation affects the model and delimits what can be modeled. This is seldom described in the literature, but lack of this knowledge often affects the ability of the reader to replicate the research. We model the expressway as a number of adjacent lanes, where each lane is divided into cells. Each cell is assumed to be 7.5 m. The most innovative aspect of our model is that we model multiple lanes as a single 1-D automaton. By extending the "Cellular Automata" (CA) paradigm to the "Global Cellular Automata" (GCA) paradigm we can represent the multilane highway with a single 1-D GCA, which can be implemented to execute faster than a traditional 2-D CA implementation and than a multi 1:D CA implementation. We present selected simulation results and outline our plan for future work.