Substack

Thursday, June 2, 2011

Optimizing transportation networks

I have blogged earlier about this (pdf here) fascinating study of unco-ordinated transportation systems by Hyejin Youn, Michael T. Gastner, Hawoong Jeong. They analyzed the 246 road links and 88 nodes of Boston's transportation network and their conclusion is striking and has significant lessons for urban transport planners,

"Uncoordinated individuals in human society pursuing their personally optimal strategies do not always achieve the social optimum, the most beneficial state to the society as a whole. Instead, strategies form Nash equilibria which are often socially suboptimal. Society, therefore, has to pay a price of anarchy for the lack of coordination among its members. Here we assess this price of anarchy by analyzing the travel times in road networks of several major cities. Our simulation shows that uncoordinated drivers possibly waste a considerable amount of their travel time. Counter intuitively, simply blocking certain streets can partially improve the traffic conditions. We analyze various complex networks and discuss the possibility of similar paradoxes in physics."


They find that contrary to conventional wisdom, the interaction of utility maximizing road users do not result in optimal traffic outcomes. Transportation flows can in reality be far from optimal even if all individuals search for the quickest paths and if complete information about the network and other users’ behaviors is available. In other words, "traffic networks can be inherently inefficient". In a game theoretic framework, they find that the Nash equilibrium (arising from self-interested road use decisions of driver) is different from the social optimum (minimized cost to society per unit transport time).

The study highlights the importance of the Braess Paradox which states that adding capacity to a network in which all the moving entities rationally seek the most efficient route can sometimes reduce the network’s overall efficiency. The graphic below highlights how certain road links add to traffic congestion due to drivers propensity to selfishly optimize on their travel times (which ironically enough, in turn results in higher travel times for the collective).



In many respects, urban transport market is a classic free-market of vehicle and road users - completely unregulated, but rife with information asymmetry (among drivers). Which routes to use and which to not use at any point in time? Which routes are best traversed through public transit as opposed to private vehicles? Which roads can be made one-way? How do we regulate vehicle flows in a road circuit?

The larger lesson it conveys to urban transport planners is about the importance of co-ordination among vehicle-users in reducing traffic congestion. Such co-ordination can be achieved by both bridging the information asymmetry between drivers and through interventions to optimize flows within transport networks.

In the prevailing "build your way out of traffic congestion" paradigm, there is very little attention paid to bridging this information asymmetry and scientifically optimizing the configuration of traffic flow networks. It is therefore not surprising that, like other unregulated free-markets, market failures (manifesting as traffic congestions) abound in this market.

1 comment:

sai prasad said...

A very relevant example comes to mind here. Some drivers try to optimise their travel by travelling opposite to traffic on a one way road (even on National Highways).

This results in suboptimal traffic conditions for society.

We have often seen the police build dividers which do not have a break for long distances. This might go against the short distance travelling folks, but certainly helps speed up traffic for most of us.

Further, this principle is pretty common in all public policy. All of us accept reasonable restrictions on our personal liberty in order to have a more lawful society.

The Boston findings have been arrived at after a lot of mathematical analysis (Which are in fact not contra intuitive).