Cellular automata for traffic simulations

Traffic phenomena such as the transition from free to congested flow, lane inversion and platoon formation can be accurately reproduced using cellular automata, Being computationally extremely efficient, they simulate large t raffic systems many times faster than real time so that predictions become feasible. A review of recent results is given. The presence of metastable s tates at the jamming transition is discussed in detail. A simple new cellul ar automaton is introduced, in which the interaction between cars is Galile i-invariant. It is shown that this type of interaction accounts for metasta ble states in a very natural way.