How car scheduling can work

Dr. Edwin R. "Chip" Kraft
Senior Associate, Reebie Associates

A recent commentary in this magazine questioned the effectiveness of current industry approaches to freight car scheduling. This echoes an earlier conclusion by Carl Martland at MIT that "the overall level of performance for the boxcar does not appear to be dramatically different in 1990 from what it was in earlier years" in spite of widespread implementation of car scheduling systems.

On this basis, one might conclude that the railroad industry's car scheduling efforts have failed and ought to be abandoned. However, today's car scheduling systems were primarily designed for (1) after-the-fact performance measurement, (2) generation of real-time alerts soon after connections have been missed, and (3) forecasting future train and yard traffic volumes. Those design objectives have been fully attained. Expecting current systems to do much more than that is just wishful thinking. Blaming car scheduling for not preventing merger-related service meltdowns is misplacing accountability. Those systems, some of which were designed over 20 years ago, never incorporated enough realism or flexibility to be usable for controlling individual car movements in real time.

The greatest weakness of car scheduling today is that the "business model" embedded in its software does not align with operational reality on the ground. That is, car scheduling systems were implemented presuming that management would commit to operation of a scheduled railroad, but that has not always been the case. Instead of being used as a vehicle to maximize adherence to schedule and maintain service reliability, car scheduling projections have often been used to develop "hit lists" of trains to annul or consolidate. In the minds of some managers at least, this role provides more benefit than the originally intended one. This also explains why service failed to improve following implementation of car scheduling systems.

Were railroads to adopt a "scheduled railroad" paradigm as Canadian National has done, the information produced by car scheduling would be much more valuable and the usefulness of such systems would no longer be the subject of debate.

It is not quite true that cars are always handled in cuts, blocks, or trains with other cars. The most important exception to this rule occurs when cars are classified in hump yards. When cars are individually sorted at the hump, routing decisions can be separately optimized for each car. Strategies for developing such classification schemes in yards can be complex, but the basic strategy is very simple. If too many cars are scheduled onto a particular outbound train, and higher priority cars are coming that absolutely must make their connections, then a sufficient number of low priority cars can be diverted at the hump either into an alternative block or into a re-hump track. An effectively managed priority system can help control costs as well as improve customer service. A priority system "levels the peaks" and "fills the valleys" of demand, utilizing train capacity that would otherwise have gone to waste, improving asset utilization, stabilizing traffic volumes throughout the network, and reducing operating cost. Properly implemented, it should yield an operational improvement rather than become an operational burden.

The real question is not improving service to low-value commodities, which are likely to continue moving by rail regardless, but whether railroads can introduce a premium carload product at a higher price into their production mix. Since most customers ship cars or trailers and not whole trains, this is the level at which traffic priorities need to be effective. Service differentiation does not improve service for all customers, but only for those who need reliable service and are willing to pay enough for it to cover any additional cost.

Since time-definite rail carload service has rarely been offered before, its marketplace potential remains to be seen. Such service is, however, what many customers have been asking for. By making more effective use of car scheduling and hump yard capabilities, railroads can offer a reliable, convenient, and less costly alternative to truck for general merchandise freight.

By enhancing car scheduling systems to take train and yard capacities into account, railroads can improve service, increase revenue, reduce cost, increase effective capacity, and improve asset utilization. This can be accomplished for a small fraction of the cost of investments otherwise needed in equipment or infrastructure to offer an equivalent level of service. Improvements to car scheduling systems and to related demand forecasting and operational control processes offer among the best investment returns now available to railroads. Those improvements should be fully funded and implementation aggressively pursued.

Dr. Kraft, who began his railroad career as a Conrail brakeman in 1978, is chairman of the INFORMS Rail Special Interest Group. Visit www.reebie.com/images/ckraft.htm.