On the Complexity of an Unregulated Traffic Crossing

TL;DR

This paper investigates the computational complexity of coordinating vehicle movements through an intersection, proving NP-completeness for the general case, and providing efficient algorithms for constrained and discrete variants.

Contribution

It establishes the NP-completeness of the traffic crossing problem and offers polynomial-time solutions for specific constrained and discrete versions.

Findings

General problem is NP-complete via 3-SAT reduction

Efficient $O(n \,\log n)$ algorithm for a constrained case

Optimal solution for the discrete version with minimal delay

Abstract

The steady development of motor vehicle technology will enable cars of the near future to assume an ever increasing role in the decision making and control of the vehicle itself. In the foreseeable future, cars will have the ability to communicate with one another in order to better coordinate their motion. This motivates a number of interesting algorithmic problems. One of the most challenging aspects of traffic coordination involves traffic intersections. In this paper we consider two formulations of a simple and fundamental geometric optimization problem involving coordinating the motion of vehicles through an intersection. We are given a set of $n$ vehicles in the plane, each modeled as a unit length line segment that moves monotonically, either horizontally or vertically, subject to a maximum speed limit. Each vehicle is described by a start and goal position and a start time and deadline. The question is whether, subject to the speed limit, there exists a collision-free motion plan so that each vehicle travels from its start position to its goal position prior to its deadline. We present three results. We begin by showing that this problem is NP-complete with a reduction from 3-SAT. Second, we consider a constrained version in which cars traveling horizontally can alter their speeds while cars traveling vertically cannot. We present a simple algorithm that solves this problem in $O(n \log n)$ time. Finally, we provide a solution to the discrete version of the problem and prove its asymptotic optimality in terms of the maximum delay of a vehicle.