Equilibrium Inverse Reinforcement Learning for Ride-hailing Vehicle Network

TL;DR

This paper introduces an equilibrium inverse reinforcement learning framework for ride-hailing networks, enabling robust driver behavior modeling and efficient policy computation in multi-agent settings with real-world data.

Contribution

It develops a novel equilibrium inverse reinforcement learning algorithm for multi-agent ride-hailing scenarios, capable of learning transferable driver reward functions and fast policy computation.

Findings

Outperforms baselines in imitation accuracy on real-world data

Computational time is independent of the number of agents

Robust to changes in supply-demand distributions and data quality

Abstract

Ubiquitous mobile computing have enabled ride-hailing services to collect vast amounts of behavioral data of riders and drivers and optimize supply and demand matching in real time. While these mobility service providers have some degree of control over the market by assigning vehicles to requests, they need to deal with the uncertainty arising from self-interested driver behavior since workers are usually free to drive when they are not assigned tasks. In this work, we formulate the problem of passenger-vehicle matching in a sparsely connected graph and proposed an algorithm to derive an equilibrium policy in a multi-agent environment. Our framework combines value iteration methods to estimate the optimal policy given expected state visitation and policy propagation to compute multi-agent state visitation frequencies. Furthermore, we developed a method to learn the driver's reward function transferable to an environment with significantly different dynamics from training data. We evaluated the robustness to changes in spatio-temporal supply-demand distributions and deterioration in data quality using a real-world taxi trajectory dataset; our approach significantly outperforms several baselines in terms of imitation accuracy. The computational time required to obtain an equilibrium policy shared by all vehicles does not depend on the number of agents, and even on the scale of real-world services, it takes only a few seconds on a single CPU.