Definition
Measures the degree to which multiple autonomous systems operating simultaneously are working in concert versus fighting each other, reported as the fraction of autonomous actions that do not trigger compensating reactions from other autonomous systems.
The Measurement Gap
When AI dosing, AI aeration, AI pump scheduling, and predictive models operate simultaneously, their individual optimizations can interact in ways no single-system metric detects. The pump scheduler creates flow variability to chase energy tariffs; the dosing system fights the variability with chemical adjustments; the aeration system compensates for changed influent patterns. Each subsystem's individual KPI looks fine: energy per million gallons improved, chemical dosing is within target, dissolved oxygen is on setpoint. The emergent behavior is invisible: increased chemical consumption to compensate for flow variability the pump scheduler created, plus process instability during transitions. No legacy metric measures whether multiple autonomous systems are cooperating or competing.
Decision Supported:
Are the plant's autonomous systems cooperating or competing, and should systems be coordinated, sequenced, or does one need to be subordinated to another?
Formula
Plain language: Define the pairs of autonomous systems that share influence over process variables. Every time System A makes an adjustment, check whether System B makes a compensating adjustment in the opposite direction on the shared variable within a defined time window. Count all compensation events across all system pairs, divide by total autonomous actions, subtract from 1.
Notation:
ASCS = 1 − ( Σ CEₐ,ᵦ / Σ Aₜ )
Variables:
| Variable | Definition | Unit |
|---|---|---|
| CEₐ,ᵦ | Compensation events between system pair (a, b): count of times system b adjusts a shared variable in the opposing direction within τ minutes of system a's action | Count |
| Aₜ | Total autonomous actions across all systems in the reporting period | Count |
| τ | Compensation detection window: maximum time lag between System A's action and System B's compensating reaction, based on process hydraulic response time | Minutes (typically 5–30) |
| (a, b) | System pairs with shared process variable influence, defined by process engineering | n/a |
Proposed formula, pending validation pilots.