Definition
Percentage of trains where the energy optimizer, network dispatcher, and crew scheduler produced mutually compatible decisions within the same planning window, measured by identifying trains where one system's decision forced another system to re-optimize.
The Measurement Gap
Railroads operate three core AI optimization systems (energy management, dispatching (network flow), and crew scheduling) that each optimize independently. No legacy metric measures whether their decisions are mutually compatible. OTP measures whether trains arrive on time but cannot see whether the on-time performance was achieved despite the three systems fighting each other: the dispatcher speeds up a train that the energy management system (EMS) slowed down, the crew scheduler assigns a crew that the dispatcher's reroute makes unavailable. Fuel consumption per GTM measures energy efficiency but cannot see whether fuel savings are being consumed by crew re-assignments triggered by energy-schedule conflicts. Crew utilization measures labor efficiency but cannot trace overtime or unplanned crew changes back to conflicts between the dispatcher's route decisions and the crew scheduler's assignment assumptions. The failure mode is three systems that each report success on their own metric while collectively destroying value through re-optimization churn: each system undoing the other's work without any metric detecting the circular waste.
Decision Supported:
Are the railroad's three core AI optimization systems working together or at cross-purposes? Should they be integrated, or should one system's output constrain the others?
Formula
Plain language: For each train in the period, check whether the EMS's energy profile was consistent with the dispatcher's speed/route plan, and whether the crew scheduler's assignment was consistent with both. Count trains where at least one system's decision forced a re-plan in another system.
Notation:
COCI = Coherent_trains / Total_trains
Variables:
| Variable | Definition | Unit |
|---|---|---|
| Coherent_trains | Trains where energy, dispatch, and crew decisions were mutually compatible: no re-optimization triggered across systems | Count |
| Total_trains | All trains operated in the measurement period | Count |
A value of 1.0 indicates perfect coherence. Lower values indicate more frequent inter-system conflicts. Track which system pair has the most conflicts: dispatch-energy, dispatch-crew, or energy-crew.
Proposed formula, pending validation pilots.