Railroads & Math
I can’t quote exactly, but a railroad executive’s statement several years ago went something like:
“ Yeah, those math guys keep pushing those computerized traffic management tools, but they just don’t realize that mainline operations are just too complex with too many exceptions to make them useful.”
Wow! What a jaw dropper. I was actually on pause for a moment as how to respond as several thoughts flashed through my mind starting with:
“So, I guess that high school algebra stuff wasn’t good for you.”
… to …
“ If it’s that complex, then how in the world is a dispatcher able to make the appropriate traffic management decisions?”
… to …
“Well if you ran your railroad in anywhere near a scheduled manner, there wouldn’t be the steady onslaught of crises.”
Fortunately, my IBM sales training some 30 years prior kicked in, and I took the CSI Miami’s Lt. Caine’s stance of tilting my head while removing (symbolically) my sunglasses, and I responded with the fail safe :
“Why do you say that?”
All that I heard for the next 10 minutes or so were examples of train movement crises that had been handled “satisfactorily” by experienced dispatchers.
Again, a series of comments went through my mind during the executive’s tyrant starting with: “How do you know they were satisfactory solutions?” … to … ‘Did the dispatcher consider options as to alternate routings given yard or crew constraints?” … to … “So, what did the dispatchers learn as to how to prevent similar situations in the future?. Again, the IBM-conditioned response:
“Really? How interesting?”
I had to get out of there; my sales training only took me to level 2, and this executive was pushing all the boundaries of cognitive rationalization.
Looking back, I can now understand where I was being too critical in evaluating this executive’s perspective of the railroad’s operations. Simply stated: He didn’t understand what he didn’t understand. He had been raised on traditional railroading based upon technologies that have changed little since their introduction in the early parts of the last century, i.e., track circuits and wireless voice. And, unfortunately the subsequent introduction of CAD platforms were limited to provide basic block status, from which dispatchers have been forced to make decisions on untimely and overly-gross data. By default, this sophistic process had become the state-of-the-art traffic management methodology.
A study of dispatcher operations performed in Sweden several years back revealed (as in “a firm grip of the obvious) that dispatchers attempt to find workable solutions instead of ideal solutions. That means, in part, that dispatchers restrict their decision making process as to handling a very, very few variables to get out of a mess. That is, they have learned through their experience to consider only a very few variables that provide the quickest solutions, but not necessarily the most cost-effective ones from the railroad’s standpoint; dispatchers are the heroes that master the inadequacies of antiquated practices based upon antiquated technologies.
In the last several years, two U.S. Class I’s have taken a more or less aggressive position that things have to change. I say more or less, because one railroad bought into the concept of Proactive Traffic Management just way too much They brought in a supplier who is so idealistic as to the movement of high speed passenger trains without the pragmatic, 80/20 perspective of what can be done with freight trains. In short, they blew it. Perhaps they are coming around, I don’t know. Reportedly, the other Class I has taken a more realistic perspective of what can be done with in-time data as to train status as to position and speed, if not just OS reports. I’m not going to give away all of the secrets here since I am a consultant that strives to make a living on dealing with the 80/20. But, permit me to just say that there are really important variables, that are much more critical than those considered by dispatchers traditionally, that can lead too much better analysis as to handling the dynamics of freight movements.
There is a next step, I can assure you, as how to merge higher-speed passenger trains into freight operation. That’s a major perspective that sets up the importance of future posts on Strategic Railroading.
Okay, so let’s test your algebra. Consider the following problem.
Train A leaves Chicago at 11:00 A.M heading East at 30 mph (instantaneous acceleration assumed) on Track 1. At 9:00 AM Train B leaves Cleveland heading West at 25 mph.on Track 2 (again, instantaneous acceleration). These are parallel tracks with the distance between Cleveland and Chicago being 350 miles. The question is where will the trains pass each other, assuming no hours-of-limits for the crews … or fuel limitations …or locomotive breakdowns … or wayside detector problems … or other disruption issues?
ANSWER: The trains meet at 200 miles West of Cleveland (150 miles East of Chicago) at 5:00 PM Eastern (clue). For the equations contact me at email@example.com.
Lastly, now take this discussion to the next level of dealing with interchange between railroads, and quickly one realizes that dispatching is one freaking mess. That is unless we use timely train status, both position and speed, regardless of the interconnecting railroad over which it is operating, fed into mathematical planners. While several railroads are beginning to realize this on their individual basis, there is clearly no industry perspective as addressed in the previous posting to this blog: Operability’s Dimensions.