An efficient transportation system is critical for the social and economic well-being of any metropolis.
However, as urbanization of the world’s population increases, all large cities have been faced with
increasing pressures on their transportations systems and the attendant scourge of congestion. In
response, cities have experimented with a multitude of approaches to address the issue, with limited
success. We propose a new, logical, doable, and fair way to rationally allocate limited road capacity
which will, if fully implemented, completely eliminate congestion.
Lately, the most commonly mentioned method to reduce congestion has been that of congestion pricing
(CP), an approach introduced by the economist William Vickrey in 1952, based on charging each
passenger a toll which equals the value of the waiting costs his presence in the system imposes on all
other users (externality costs). If such tolls, which naturally vary over time, can be computed and
charged, the resulting congestion will be reduced, by the economist’s invisible hand, to its “socially
optimal level”.
In spite of the intuitive appeal of this approach, it suffers from a few weaknesses. In the first place, even
if the “optimum” could be achieved, the resulting congestion would not be completely eliminated. Rather,
the best we can hope for under the scheme is some reduction in congestion. But even that reduction is not
likely: The computation of the optimum tolls depends on specific modeling assumptions and is quite
involved. And even if those tolls could be precisely computed, their complex dynamic pattern of
variation makes their imposition on the public very tricky. In practice, the “optimal” CP tolls are
approximated by a simple two-tier tolls system (“peak hour tolls”) which is determined by trial-and- error
or guess-work and which may be far from optimal. Thus, the benefits of CP may be quite small, and are
often transitory and fleeting. Indeed, there is a wide spread perception among travelers that the real
motivation behind CP systems is not the alleviation of congestion but rather the pursuit of additional
municipal revenues; that people will pay the tolls but will still be stuck in traffic. Thus, the method is
often politically unpopular.
Currently none of the approaches proposed, including Congestion Pricing, actually manages traffic and
none will completely eliminates congestion. We propose a new reservation-based approach for managing
traffic which is based on principles from Operation Management, Network Flow Theory and Just in Time
Scheduling Systems. The basic observation is that, in a capacitated system, any load on the system
beyond the capacity of its bottleneck(s) is a dead-weight loss: it increases service times but does NOT
contribute to any increase in the rate at which service is produced or provided. Thus, rather than using
pricing schemes to reduce usage indirectly, our method utilizes a Routing, Scheduling and Reservation
System (RSRS) to route, schedule and reserve all trips to ensure that the volume of traffic on any road
segment does not exceed capacity at any time. To get access to the road, one would need to reserve a
specific route at a specific time. A companion Vehicle Monitoring System (VMS) is utilized to ensure
that legitimately reserved trips are in fact recognized and admitted into the network.
The approach requires that, at periods of high demand, passengers forfeit their right to initiate trips at will.
This feature will, no doubt, raise significant resistance. It is important to remember, however, that while
passengers’ start-of-trip may be delayed in accordance with the availability of road capacities, their
arrival time at the final destinations will not: In a way, the reservation system “preserves one’s position
in the line” as if she was actually on the road, stuck in traffic.
The principles involved in operating a reservation system are well known and widely used. Restaurants,
physician offices, golf courses, tennis courts and numerous other capacity constrained outfits, both in the
private and in the public sectors, routinely use a reservation systems to manage limited capacity. The Air
Traffic Control System, which often delays the takeoff of a plane in case of air travel congestion on route
or at destination, is perhaps the closest to what we have in mind. Nevertheless, the method was hitherto
not considered in the context of urban traffic due to the vast scale and complexity of the systems
involved.
However, with the recent dramatic advances in communication technologies, computational capacity and
sensory systems, all the elements needed to implement our approach are now in place.
Our method offers many advantages over CP systems. In the first place, since reservations are driven by
capacity and not by pricing, the method has the potential to completely eliminate, rather than reduce,
predictable congestion, while maximizing the utilization of existing transportation infrastructure. The
method is also better equipped to deal with scheduled congestion-inducing events such as concerts and
sporting events, as well as with unpredicted traffic disruptions such as those caused by accidents. It can
be easily extended to accommodate situations involving limited parking availability, and is perfectly
positioned to accommodate future scenarios in which a large number of self-driving cars are predicted to
enter the urban scene. Finally, while full benefits require full implementation, the method can be easily
implemented on a partial basis, operating, say at specific time periods or covering a restricted area.
And, most critically, the method allows for great flexibility in deciding how reservations are allocated (as
opposed to how many). Some spots could be sold or auctioned, using a variety of pricing schemes. At
the same time, some fraction of the network capacity can be allocated with an eye to promoting other
municipal priorities such as inspiring a sense of fairness or social justice, supporting civil services, and
allowing the city’s various business sectors to prosper. As an example of the later point, the system could
prioritize , for instance, travelers heading for an airport to catch specific flights. Delivery trucks, on the
other hand, could be directed to non-congested time periods.
It noteworthy that our method allows for two potential streams of revenues. In addition to charging
directly for (some or all) reservations, one can also include charges or fines for anyone who uses the
system without reservation thereby creating congestion and delays for those who have complied with the
scheduling system.
The main point is the framework is flexible and can accommodate a rich set of implementation options.
This flexibility, in turn, allows each city to strike a unique customized balance among several competing
municipal and political objectives.