MEL Interaction Risk Score (MIRS)

GMI-AVI-041AviationResearch-Forward

Definition

The ratio of the interaction-adjusted dispatch risk (accounting for failure mode dependencies between concurrent MEL deferrals) to the independent-sum dispatch risk (the AI's assumption that concurrent deferrals are probabilistically independent). A score near zero means the AI's independence assumption is safe. A score above 0.25 means the AI is materially underestimating dispatch risk by ignoring how deferred items interact.

The Measurement Gap

The MEL is designed to permit dispatch with one known inoperative system at a time. Regulatory probability assumptions built into dispatch risk calculations treat each deferral as independent: the probability that a second failure occurs is not assumed to increase because a first failure is already present. AI MEL management systems inherit this assumption and compound it: when multiple items are deferred simultaneously, the AI evaluates each against its individual MEL provisions and concludes the combination is acceptable if each individual item is acceptable.

The problem is that failure modes interact. A deferred weather radar combined with a deferred terrain awareness warning system backup mode and a deferred autopilot approach mode do not create three independent risk increments: they create a specific operational scenario where the pilot's situational awareness and workload management tools are degraded simultaneously. Engineering interaction rules exist in some maintenance engineering organizations, but they are not systematically integrated into AI MEL management algorithms. MIRS makes the gap visible: it requires organizations to quantify the difference between the risk the AI computed and the risk that accounts for interactions. Because developing interaction rules is itself a research task, this is classified as Research-Forward: the metric defines what to measure and provides the formula, but the interaction rules data must be developed as a separate engineering program before MIRS can be calculated.

Formula

Plain language: Subtract the AI's independent-sum risk estimate from the interaction-adjusted risk estimate, then divide by the independent-sum estimate to produce the ratio of additional risk that the AI's independence assumption is missing.

Notation:
MIRS = (R_interaction − R_independent) / R_independent

Variables:

Variable Definition Source
R_interaction Dispatch risk estimate accounting for failure mode dependencies between concurrent deferrals Engineering interaction rules model
R_independent Dispatch risk estimate treating each concurrent deferral as probabilistically independent AI MEL management system output

Interaction adjustment method: R_interaction requires an engineering interaction rules matrix that specifies, for each pair of MEL items that may be co-deferred, whether the failure modes are: (a) independent, (b) correlated (shared failure source increases both), or (c) operationally compounding (one failure increases workload/consequences if the other occurs). For correlated and compounding pairs, joint probability adjustments are applied to the combined risk calculation.

Proposed formula, pending validation pilots.

At a Glance

Category
Operational Decision Quality Under Pressure
← MEL Deferral Operational Context Score (MDOCS)Demand Forecast Strategy Alignment Index (DFSAI) →

Assess Your Organization

The AI-GPS Assessment evaluates your AI governance maturity and recommends KPIs like this one for your implementation roadmap.

Take the Free AI-GPS Assessment