Your Cart

Your cart is empty

Home  >  Volume 52

17. A DYNAMIC ROAD TRAFFIC ROUTING MODEL USING Q LEARNING AND MARKOV DECISION PROCESS by O. A. Sanya and O.A. Bello Volume 52 (July & Sept., 2019 Issue)
Sale price: $0.00
Quantity:

A DYNAMIC ROAD TRAFFIC ROUTING MODEL USING Q LEARNING AND MARKOV DECISION PROCESS

O. A. Sanya and O.A. Bello

Department of Mathematical and Physical Sciences, College of Sciences, Afe Babalola  

University, Ado Ekiti, Ekiti State, Nigeria

Abstract

Road traffic delay has become a menacing challenge in cities around the world. This research work gives an overview of a dynamic road traffic routing system that can help in determining optimized path from one point to another in a given road network in terms of path and the associated delay. This is achieved by simulating real-world traffic situations on selected road paths in the University of Lagos. Q Learning, a reinforcement learning algorithm was used to facilitate an efficient and dynamic handling of parameters. Q Learning is even more appealing because it inherently has routing capability and it easily allows us to train an agent that represents a road user to make optimal decision. The result indicated that the system will always select a route that is optimal in relation to prescribed constraints. Markov Decision Process (MDP) was adopted in modeling the research

Keywords: Routing, Q – Learning, Reinforcement Learning, Markov Decision Process.

Close
Loading...