An Improvement to Flock Traffic Navigation Algorithm Using Linear Regression Techniques

Flock traffic navigation (FTN) is a recently approach to solve vehicular traffic congestions in cities using multiagent technology. These agents are instructed to cooperate with others in order to form flocks, which gain a speed bonus, therefore improving individual travel time. Identifying partners is a critical issue, since it is a time consuming process. In this paper we present an improvement to the negotiation algorithm using linear regression techniques in order to decide which subset of agents could be the best possible partners. Finally, we also present early experimental data that support these ideas.