Simulação microscópica distribuída de tráfego

Abstract

Most of the major cities of the world face traffic problems. Simulation is becoming one of the most used tools in the search of solutions in traffic. Simulation uses mathematical techniques, computer and statistics to represent traffic. The simulation allows the analysis of possible traffic configurations to be implemented to improve the flow of vehicles in a certain place. Models of simulation of traffic can be classified in many ways. The most used is about the level of detail with which the models represent the real world and the way the behavior of vehicles is modeled. Following this classification, the models can be macro, meso or microscopic. Microscopic models simulate the system entities individually with a high level of detail. Each pair vehicle-driver has an own behavior and is an active component in the simulation. Not only vehicles and drivers are modeled, but any other entity that can influence the traffic, like traffic controllers, specific roadways and incidents. Modeling traffic in the microscopic level today represents the state of the art for traffic simulators. The more complete the model, the more realistic will be the simulation. On the other hand, it needs a great computational effort. Some simulators require supercomputers or high performance computers and thus have a high cost. The alternative to the supercomputers is the use of distributed computing. A distributed environment allows the simulation program to run on multiple processors in parallel and thus gains in performance. Using a distributed environment requires a strategy of division of tasks between processors and a solution to synchronize these tasks. This work aims to identify and implement a solution for a simulator microscopic traffic in a distributed environment, providing a low-cost alternative for study and improvement of traffic.