A Multiagent Simulation for Traffic Flow Management with Evolutionary Optimization

A traffic flow is one of the main transportation issues in nowadays industrialized agglomerations. Configuration of traffic lights is among the key aspects in traffic flow management. This paper proposes an evolutionary optimization tool that utilizes multiagent simulator in order to obtain accurate model. Even though more detailed studies are still necessary, a preliminary research gives an expectation for promising results. 1 Traffic flow management Early models of traffic flow (called macroscope models) treated vehicles in the collective manner basing on the analogy to particles in a fluid [7]. Later on, more precise (mezoscope) models were being created consecutively (mostly based on gas kinetics) [5]. Currently, a multiagent simulation mechanism provides much more efficient microscope model where each vehicle can be regarded separately allowing for highly detailed analysis including collision avoidance [4], traffic virtualization [2], interactions with pedestrians [3], etc. Traffic flow management varies from tracing main roads average capacity across certain area (where macroscope models give satisfactory results) to very low-level manipulations including re-arrangement of lanes, modifying traffic lights configuration, planning bridge locations, etc. where microscope models are most suitable [9]. 2 Multiagent traffic flow simulator The agent-based traffic flow simulator described in [9] is the universal microscope model. The environment for agents in this model comprises of a system of city streets defined in XML files using the following entities illustrated in Figure 1: