Delays and traffic bottlenecks represent a significant economic loss and inconvenient issues in transportation of persons and goods. That is, transportation systems still have a wide spectrum of issues that need to be addressed. In this sense, Intelligent Transportation Systems (ITS) are expected to provide an efficient and modern way to deal with these issues. Nevertheless, underlying technologies are not always reachable for urban settlements with budget restraint such as in rural areas and in developing countries. In this paper, we present a novel and affordable architecture aiming to improve vehicular traffic management based on intelligent Virtual Police Agents (VPA). Our architecture is inspired from real police agents figure which are oriented to manage traffic issues efficiently. The replication of basic features of police agent role in VPA results in collaborative and hierarchical agent-based relationship architecture. This framework aims to establish system-optimal based in traffic flow equilibrium with maximum-flow and minimum-cost.