Degree Name

Master of Computer Science - Research


School of Computer Science and Software Engineering


This thesis reports on the development of a multi-agent approach to distributed traffic optimisation. In particular, I propose a solution to the dynamic traffic assignment problem in a decentralised manner and then I introduce the new infrastructurelessly decentralised traffic information system. By using this system, each vehicle agent is able to update the current traffic condition through vehicle-to-vehicle communication. For solving dynamic traffic assignment problem, I propose a novel completely decentralised multi-agent coordination algorithm, which is a synergy between dynamic distributed constraint optimisation problem (DynDCOP) algorithm and auction. Using this algorithm, vehicle agent is able to reduce its individual travel time as well as total travel time of overall system. This simulation is carried out in order to evaluate different traffic planning algorithms that include decentralised uncoordination, centralised coordination and decentralised coordination algorithms. Finally, the experimental results show that the performance of proposed decentalised coorindation algorithm is high in comparison to centralised coordination algorithm.