Self-adaptive and multi-agent reinforcement learning in route guidance system

Several challenges in traffic control in route guidance system causes increasing number of vehicles to transport goods and people in our society. The concept of autonomous agents fits most actors in transportation systems: the traffic, the expert, the driver. More so, traffic signals and intersection can also be regarded as an autonomous agent. Though, there are increased number of agents, typical agents make response to changes in their environment and are highly self-adaptive, but create an unpredictable collective pattern, and response in a highly coupled environment, most challenges for standard techniques are created by this domain in route guidance system from multi-agent systems such as reinforcement learning and self-adaptive. This research has two main goals in route guidance system: first, to present problems, methods, new approaches; and second, open problems and challenges are highlighted so that future research in route guidance system using multi-agent systems will be able to address them.