Synchronizing Multi-Threaded Activity under Pressure: A Simulation Study of Contingency Response in Envisioned Air Traffic Management Operations

Technical Summary: Synchronizing Multi-Threaded Activity under Pressure

Problem Statement
The paper addresses the challenge of synchronizing interdependent activities within envisioned Advanced Air Mobility (AAM) and Urban Air Mobility (UAM) operations, specifically during contingency events such as unexpected aerodrome reconfigurations. While existing systems engineering approaches often rely on static models of work or positive test strategies to verify core functions, these methods frequently fail to uncover synchronization requirements arising from the temporal dynamics of distributed work. In complex, high-tempo environments involving both human and automated actors (e.g., pilots, fleet operators, service providers), coordination breakdowns can occur when agents act on inconsistent information or fail to align with the pace of environmental changes (e.g., wind shifts). The study posits that without explicit consideration of work dynamics and temporal behavior, designs may lead to "locally adaptive, globally maladaptive" outcomes, such as race conditions, overtaking, or system saturation.

Methodology
The authors employ a computational modeling and simulation approach based on a three-phase framework previously established to explore trade-offs in envisioned systems. The study focuses on a specific contingency: an unplanned aerodrome reconfiguration at a vertiport located in a constrained corridor near Dallas Fort-Worth (DFW), triggered by a sudden wind shift that renders the current approach direction unsafe.

• Scenario: Five aircraft are en route to the vertiport. Upon a wind shift, the first aircraft executes a go-around, necessitating that subsequent aircraft reroute and re-negotiate landing directions.
• Variables: The simulation explores a trade space defined by four key independent variables:
  1. Decision-Making Authority: Centralized (a Provider of Services [PSU] dictates rerouting) vs. Decentralized (pilots/fleet operators negotiate new Operational Intents [OIs] autonomously).
  2. Wind Change Timing: When the disruption occurs relative to the arrival of the aircraft.
  3. Pilot Communication Lag: The delay between a pilot deciding to go-around and the digital system receiving the reason and intent (0s vs. 30s).
  4. Between-Aircraft Interval (Buffer): The minimum time separation required between aircraft at waypoints (15s, 30s, 45s, 60s).
• Follow-up Analysis: A secondary set of simulations varied the Wind Report Generation Frequency (5s, 7s, 10s, 20s) to investigate interactions between communication lag and data update rates, specifically within the decentralized architecture.
• Metrics: Performance was measured by total mission completion time (time until the last aircraft lands), the number of OI negotiations, the number of go-arounds, and network analysis metrics (modularity and transitivity) of information exchange.

Key Results
The simulation yielded several counter-intuitive findings and identified specific synchronization patterns:

1. The "Lag" Paradox: Contrary to the assumption that instantaneous communication is always superior, the study found that a 30-second communication lag sometimes resulted in faster mission completion times than zero lag. In "Lag 0" scenarios, automated systems often triggered new OI requests before wind reports were updated, leading to decisions based on stale data. This caused the system to reject valid routes or fail to reroute, eventually requiring more time to resolve non-conformance. The 30-second lag inadvertently provided the necessary temporal slack for wind reports to update before automated decisions were finalized.
2. Architecture Performance: The centralized architecture generally outperformed the decentralized one, exhibiting faster completion times and fewer OI negotiations. The decentralized architecture suffered from increased negotiation conflicts and "race conditions," where aircraft competed for optimal slots, potentially violating safe separation distances and landing order.
3. Three Synchronization Patterns: The follow-up analysis revealed three distinct patterns describing how communication lag and wind report frequency interact to affect reconfiguration time:
   • Pattern 1: Lag increases reconfiguration time when wind reports update after the first go-around but before system-level monitoring.
   • Pattern 2: Lag increases reconfiguration time when wind reports update before any go-arounds (allowing immediate, correct OI creation without lag).
   • Pattern 3: Lag decreases reconfiguration time when the delay prevents a premature OI proposal that would otherwise conflict with a subsequent aircraft's proposal, thereby avoiding an additional go-around.
4. Critical Timing Requirement: The study identified a critical synchronization requirement: the wind report generation frequency must be higher (i.e., faster updates) than the conformance monitoring interval. If the wind report interval exceeds the monitoring interval, the system may fail to detect non-conformance or trigger rerouting because it lacks updated weather data, leading to a state of unresolved non-conformance.

Significance and Claims
The paper claims that work-based modeling and simulation serve as a vital discovery tool for identifying synchronization requirements early in the systems engineering cycle, prior to prototyping. Its primary contribution is demonstrating that:

• Static models are insufficient: They fail to capture the emergent properties of distributed work, such as race conditions and the negative impacts of acting on stale data.
• Synchronization is temporal: Effective coordination requires managing the "time-costs" of information propagation. Instantaneous communication is not inherently superior; in some cases, it creates asynchrony by allowing agents to act on inconsistent versions of reality.
• Design Implications: Future AAM systems must incorporate mechanisms that verify information consistency before executing interdependent work. Specifically, automated rerouting services should wait for updated weather reports, and decentralized systems require explicit strategies to maintain pseudo-sequencing to avoid dangerous race conditions.
• Role of Simulation: The study validates the use of computational work models to "stress" envisioned worlds, uncovering hidden dependencies and trade-offs that traditional compliance testing misses. It emphasizes that the design of cooperative traffic management must account for dynamic shifts in architecture and the specific timing of information updates to ensure resilience.

The authors conclude that these findings provide the grounding necessary to inform technical trade-offs and influence system design decisions, moving beyond simple automation unit testing to a more holistic understanding of system synchronization.