A unified framework of proactive self-learning dynamic pricing for high-occupancy/toll lanes

This article presents a unified framework to determine dynamic pricing strategies for high-occupancy/toll (HOT) lanes. The framework consists of two critical steps, system inference and toll optimisation. The first step is to mine traffic data in a real time manner to learn motorists' willingness-to-pay, estimate traffic state and predict short-term traffic demand. The attained knowledge is then used in the second step to explicitly optimise toll rates for the next rolling horizon to maximise the freeway throughput while ensuring a free-flow travel speed on HOT lanes. This article discusses the details of each step and how to implement them. The framework is validated in a simulation environment based on a multi-lane hybrid cell transmission model. It is demonstrated that the framework is efficient, effective and flexible, and has the potential to be readily implemented in practice.