Abstract
Airplane boarding process is an example where disorder properties of the system are relevant to the emergence of universality classes. Based on a simple model, we present a systematic analysis of finite-size effects in boarding time, and propose a comprehensive view of the role of sequential disorder in the scaling behavior of boarding time against the plane size. Using numerical simulations and mathematical arguments, we find how the scaling behavior depends on the number of seat columns and the range of sequential disorder. Our results show that new scaling exponents can arise as disorder is localized to varying extents.
Introduction
Airplane boarding is more than just a logistical challenge; it is a complex statistical physics problem. A central question in this field is how the average boarding time $\langle T \rangle$ scales as the number of rows $N$ increases. Previous research has yielded conflicting results, with some models predicting a square-root scaling of $N^{1/2}$ and others suggesting a steeper $N^{0.69}$.
Key Takeaway: This study resolves these contradictions by examining sequential disorder—the degree to which the passenger entry sequence deviates from a perfect row-by-row order. The authors demonstrate that for multi-column aircraft, strict row-by-row boarding (localized disorder) is actually less efficient than random boarding (global disorder) because it forces passengers to cluster in the aisle, leading to linear scaling ($\langle T \rangle \sim N^1$).
Modeling Sequential Disorder
To investigate these effects, the authors define different regimes of passenger arrival sequences. "Localized disorder" occurs when passengers are randomized only within small groups of $m$ rows, while "globalized disorder" involves randomization across the entire aircraft.
The study employs Gaussian shuffling to simulate arrival time fluctuations and uniform shuffling to analyze the coexistence of local and global disorder. By mapping the boarding process to the mathematical "longest increasing subsequence" problem, the authors can predict scaling exponents based on the length of the longest chain of passengers who block one another.
Methods and Numerical Simulations
The research utilizes a discrete 1D model where an airplane has $N$ rows and $c$ columns. Passengers enter sequentially and are blocked if a passenger ahead is moving or stowing luggage. The authors conducted extensive numerical simulations for plane sizes ranging from $N=10^2$ to $N=10^5$, averaging results over at least 100 samples to ensure statistical significance. They calculated the effective scaling exponent $\bar{z}$ to determine the asymptotic behavior of the system.
Scaling Regimes and Universality Classes
The analysis reveals three distinct asymptotic scaling regimes for boarding time:
- Globalized Disorder: $\langle T \rangle \sim N^{1/2}$. This represents the universal behavior for random boarding, where passengers are spread out and less likely to block each other.
- Localized Disorder (Single Column): $\langle T \rangle \sim N^0$. In single-column planes with localized disorder, boarding time saturates or grows only logarithmically.
- Localized Disorder (Multi-Column): $\langle T \rangle \sim N^1$. This is the least efficient regime, where boarding time grows linearly with plane size due to high aisle congestion.
Interestingly, the authors found a "vacancy threshold" at $\phi_c = 1 - 1/c$. At this critical point of partial occupancy, the boarding time scales as $N^{1/2}$ regardless of the disorder type. Furthermore, even a small amount of global disorder (e.g., a few unpunctual passengers) can shift a localized system back into the more efficient $N^{1/2}$ regime.
Conclusion and Future Directions
This research provides a comprehensive view of how sequential disorder dictates the efficiency of airplane boarding. It highlights the counter-intuitive benefit of "chaos" in multi-column aircraft, where global randomization outperforms strict row-wise control. While the model uses simplifications like synchronous updates and uniform sitting times, it establishes a robust framework for understanding universality classes in boarding processes. Future work may incorporate stochastic sitting times and seat interference effects to further refine these scaling laws.
Summary updated