As a former railroader very much involved in operations improvement, I'm not motivated to a nice response to that one. We were trying to do all of the above in the 60's. It was much more difficult then because we had only primitive computer power.
At the time of the PC merger, we were actually running a daily report that showed movement of auto rack cars compared with schedule, looking at individual movements. We had begun research work on a car scheduling system. After the
PC merger, the VP in charge of that went to the MP.
Railroads have been working on the car scheduling problem since then. Such a system was being implemented on the MP when it was acquired by the UP and then installed over the whole UP system.
There is one facotr that makes it close to impossible to actually realize precision scheduling of individual carload moves. That is the unpredictable pattern of inputs to the system. Shipperes have the option of releasing an individual car shipment when it's convenient for them. Another is random breakdowns that delay cars and cannot be avoided at any reasonable cost. You can get zero delay only by spending an infinite amount of money. It's not like airplances and space vehicles where there is much risk of minor flaws killing people.
