TLDR: A new continuous-time mixed-integer linear programming (MILP) model is introduced for integrated aircraft hangar scheduling and layout. This model efficiently solves complex spatio-temporal problems, overcoming scalability issues of traditional discrete-time methods. It consistently outperforms heuristic approaches, offering significant economic benefits and providing valuable insights for managers, especially in handling congested schedules, and is supported by a visualization tool for practical application.
Managing aircraft maintenance hangars efficiently is a major challenge in the multi-billion dollar global aircraft maintenance, repair, and overhaul (MRO) industry. These hangars are often bottlenecks in the MRO process, making it crucial to optimize their space and time usage. This involves complex decisions about when to bring aircraft in and out, and where to park them, as these choices directly impact fleet readiness and financial performance.
Historically, solving this complex problem has relied on exact methods using discrete-time formulations. While some advanced models have shown promise, they often struggle with scalability, especially when dealing with a high number of aircraft or highly congested arrival times. For instance, some models failed to find optimal solutions for even a small number of aircraft within reasonable time limits if arrival times were too close together. This limitation often forces the use of simpler, faster methods called heuristics, which unfortunately sacrifice the guarantee of finding the best possible solution.
A new research paper, An Efficient Continuous-Time MILP for Integrated Aircraft Hangar Scheduling and Layout, introduces a novel approach to overcome these limitations. The paper proposes a continuous-time mixed-integer linear programming (MILP) model. By treating time as a continuous variable, this new formulation significantly reduces the model’s complexity, including the number of variables and constraints, compared to its discrete-time predecessors. This allows for the efficient solution of larger and more realistic problem instances.
Key Contributions of the New Model
The research highlights three main contributions. First, the novel continuous-time formulation itself, which fundamentally changes how time is handled in the optimization problem. Second, a comprehensive computational study that benchmarks the exact model against a fast, priority-rule-based heuristic. This comparison quantifies the substantial economic benefits of achieving an optimal solution. Third, the development of an insight-driven decision support framework, which includes a custom-built visualization tool. This tool transforms the model’s numerical outputs into an interactive dashboard, helping managers understand the optimal plan and the reasoning behind complex strategic decisions.
How It Works
The model considers the hangar as a rectangular area with a single entrance/exit. It accounts for aircraft dimensions, requires safety buffers between aircraft and walls, and ensures that no two movement events (aircraft rolling in or out) occur simultaneously. Once an aircraft is parked, its position is fixed until departure, simplifying the model while aligning with operational practices where repositioning is costly and time-consuming.
The objective of the model is to minimize total operational costs. These costs include penalties for rejecting new aircraft requests, and penalties for delays in aircraft arrivals and departures. There’s also a small cost component that encourages a compact and organized spatial layout by placing new aircraft closer to the hangar’s origin.
Performance and Insights
The computational results are compelling. The MILP model consistently finds provably optimal solutions for small to medium-sized instances (up to 25 aircraft) in mere seconds. For larger cases (up to 40 aircraft), it delivers high-quality solutions within known optimality gaps, even if finding the absolute optimum takes a few hours. This is a significant improvement over older discrete-time models that struggled with similar sizes.
When compared to the Automated Constructive Heuristic (ACH), the exact model demonstrates significant cost savings. The performance gap between the optimal MILP solutions and the heuristic solutions widens as the problem size increases, underscoring the substantial economic value of optimal planning over simpler, greedy approaches.
A crucial finding is the model’s robustness to event congestion. Unlike event-based models that slow down dramatically when arrival times are tightly packed, this continuous-time formulation’s solution time remains stable, primarily dependent on the number of aircraft rather than the density of their schedules. This makes it a more reliable and scalable solution for dynamic, real-world environments.
Also Read:
- Optimizing Electric Vehicle Deliveries: A New Framework for Smarter Routing and Charging
- Unlocking Advanced Scheduling: A New Approach to Cumulative Functions in Open-Source Solvers
Practical Applications for Managers
The research offers several actionable insights for hangar managers. The MILP model can make counter-intuitive but globally optimal decisions, such as rejecting a high-priority aircraft or strategically delaying another, if it prevents more costly conflicts overall. This ability to identify system-wide trade-offs is difficult for human planners or simple heuristics.
Furthermore, the model provides strategic flexibility. Managers don’t have to choose between a fast, suboptimal heuristic and a slow, optimal model. They can decide based on their operational context: for long-term strategic planning, they can run the solver to full optimality. For daily tactical decisions, they can set an acceptable optimality gap (e.g., 5% or 10%) to get a high-quality solution much faster, with a known quality guarantee.
To enhance practical application, a custom visualization tool was developed. This interactive dashboard translates the complex numerical solution into an intuitive graphical interface, allowing managers to see the hangar’s state at any time, anticipate bottlenecks, and understand the logic behind the optimal plan. This bridges the gap between abstract models and tangible operational plans.
This research provides a powerful framework for optimizing hangar operations, leading to improved throughput, reduced costs, and valuable managerial insights, paving the way for more efficient aircraft maintenance globally.
