Supervised learning from human performance at the computationally hard problem of optimal traffic signal control on a network of junctions.

Simon Box

doi:10.1098/rsos.140211

Abstract

Optimal switching of traffic lights on a network of junctions is a computationally intractable problem. In this research, road traffic networks containing signallized junctions are simulated. A computer game interface is used to enable a human ‘player’ to control the traffic light settings on the junctions within the simulation. A supervised learning approach, based on simple neural network classifiers can be used to capture human player's strategies in the game and thus develop a human-trained machine control (HuTMaC) system that approaches human levels of performance. Experiments conducted within the simulation compare the performance of HuTMaC to two well-established traffic-responsive control systems that are widely deployed in the developed world and also to a temporal difference learning-based control method. In all experiments, HuTMaC outperforms the other control methods in terms of average delay and variance over delay. The conclusion is that these results add weight to the suggestion that HuTMaC may be a viable alternative, or supplemental method, to approximate optimization for some practical engineering control problems where the optimal strategy is computationally intractable.

Highlights

The set of tasks at which humans can outperform machines has been steadily shrinking. This progress has been punctuated by landmark events where a machine is shown to be able to match or exceed human performance at a task that was previously only routinely performed by humans; for example, driving a car in urban traffic [1], playing backgammon [2] or competing in a
We evaluate the performance of signal control strategies by looking at both mean delay— μ(θ ) and standard deviation over delay—σ (θ ), which is a good proxy for how equitable the treatment of vehicles is under a signal control strategy
The performance of human-trained machine control (HuTMaC) has been demonstrated on two road network models, where it exhibited comparable performance to the benchmark control systems of MOVA, SCOOT and temporal difference (TD) control

Summary

Introduction

The set of tasks at which humans can outperform machines has been steadily shrinking. This progress has been punctuated by landmark events where a machine is shown to be able to match or exceed human performance at a task that was previously only routinely performed by humans; for example, driving a car in urban traffic [1], playing backgammon [2] or competing in a. There may be a practical way to employ supervised learning from expert human performance as a method to improve machine performance at these tasks. This paper proposes that this is the case for the task of traffic light signal control on a network of road junctions

Methods

Results

Discussion

Conclusion