Real-time Decision-support Models For Hub Control Centers
New turnaround-management and payload-management decision-support models address airlines’ needs to make real-time decisions that optimize aircraft turn times and more efficiently transfer passengers, luggage and crew.
Real-Time HCC Models
In a past article of Ascend (2012, Issue No. 1, “Common Ground: In The Air”), the author, Sergey Shebalov, explains how to bring together an operations control center (OCC) and a hub control center (HCC) through common IT support. Based on that work, the operations research team at Sabre Airline Solutions® has developed advanced decision-support models as a prototype that addresses airlines’ needs to make timely decisions related to turnaround management and payload management.
A new turnaround-management model optimizes the schedule for tasks required to turn around an aircraft at the hub for its next flight. Likewise, the new payload-management model optimizes transfers of passengers, luggage and crews.
These two models, developed by the operations research team at Sabre Airline Solutions, demonstrate that their capabilities can be leveraged to provide real-time decision support based on computational results.
The OCC and HCC develop their own operational plans for an airline’s daily operations.
The OCC focuses on an airline’s overall network management. It is fully responsible for short-term planning, tracking and recovering the flight and crew schedule, as well as tail/maintenance assignment.
The HCC is responsible for ground tasks at a hub station, such as reallocating the gate, reassigning ground tasks and reaccommodating passengers. It can also suggest limited changes of the flight schedule (e.g., delay a flight for 10 minutes). If the OCC’s decisions overlap those of the HCC, the HCC communicates suggested changes to ensure both groups are aligned. This level of communication is essential to solve disruptions to daily operations.
HCC Decision-Support System
Airlines have spent years building decision-support systems for OCCs. Now, it has become extremely important for airlines to review the overall hub management function to improve passenger experiences (e.g., lowering the chance of delays and stranding passengers at an airport) and reduce costs, such as those caused by delays and passenger reaccommodation.
The ideal HCC decision-support system consists of several functional modules, such as schedule management, turnaround management, payload management and resource management. When implementing such a system, airlines can take advantage of their existing systems. Many airlines already have tools for schedule management and resource management (e.g., staff- and gate-planning systems). Therefore, they can focus more intently on turnaround management and payload management and build individual decision-support models for each functional area.
Both models can work in a hub-control environment, so multiple tasks can be automatically generated for resource management. In addition, focusing on individual modules just one time provides an evolutionary way for airlines to build their integrated HCC systems. It is less risky and requires a relatively small investment.
The OCC And The HCC Must Communicate
Instantaneous information exchange between an airline’s operations control center and its hub control center is essential. In daily operations, both OCC operations data updates, such as flight schedule changes, and the HCC tracking system can trigger violations and warning messages (e.g., passenger misconnections, ground tasks running overtime, etc.), which drive the HCC decision-support system to fix violations by accessing all required information. Once that occurs, changes by the decision-support system will feed back to the OCC.
Turnaround management focuses on tasks and processes associated with the turn of an aircraft. Aircraft turnaround starts when an aircraft arrives at the gate (chocks-on moment) and ends when the aircraft leaves the gate (chocks-off moment). Tasks must be completed between the two moments and follow a specific order. For example, passenger boarding cannot start if passenger disembarking is still in progress.
The turnaround-management decision-support model helps ground-service coordinators at the HCC intervene when a disruption occurs during the process of a turnaround. It can find the optimal way to minimize the total cost of the turnaround and meet the turnaround duration by rearranging tasks, such as cancelling a specific task or speeding up a task (reducing the time to complete the task).
The model determines the start and end time of each task, and it recognizes critical tasks as those that cannot be delayed without causing further disruption. Based on this, ground-service coordinators at the HCC can assign additional staff, as well as reschedule and restrict task sequences.
The decision-support model enumerates all possible options for each task, and each option for one task corresponds to one combination of start times and end times. For example, cabin cleaning can begin at 10 a.m. and end at 10:15 a.m., or start at 10:05 a.m. and end at 10:15 a.m. There is a cost associated with each option. The model, therefore, choses one option for each task and minimizes the total cost of the turnaround.
To find a solution quickly, the model only considers reasonable options for each task. In practice, the discrete start time and duration can be counted in distinct time units, such as minutes rather than seconds, which can reduce the number of options. Meanwhile, by establishing a precedence relationship between tasks, the airline can obtain the earliest start time and latest end time of each task and further reduce the number of options. Moreover, it can design turnaround templates, such as a small set of feasible turnaround durations like 25, 35 or 45 minutes, for one aircraft to restrict the number of feasible options.
The operations research team at Sabre Airline Solutions tested its turnaround-management model with one example having 51 tasks. Each task was allowed to speed up to 50 percent of its originally planned duration. The decision-support model considered thousands of options for all tasks and found the optimal solution in less than one second. This level of real-time decision support helps an airline quickly determine whether or not it is feasible to make a shorter turnaround and how to rearrange tasks.
Four Core Modules
The HCC decision-support system consists of four modules. Schedule management is responsible for tracking flight operations during the day and reacting to disruptions. Turnaround management focuses on tasks and processes associated with turning an aircraft. Payload management handles passengers, luggage and cargo connections violations. Resource management works in a hub-control environment so multiple tasks can be automatically generated for airport resources.
Payload management handles entity (i.e., passengers, baggage, crew and tail) connections. The payload-management decision-support model helps HCC controllers fix connection violations by automatically reviewing multiple recovery options.
For example, if a particular passenger connection is disrupted due to late arrival of an inbound flight, the model will check whether fast transfer or ramp transfer options are available. If these options do not resolve the issue, it will evaluate recovery options such as an outbound flight delay within threshold, reaccommodation on a later host flight, reaccommodation on another airline’s flight or an overnight stay. The total cost for all these options is considered, and the best solution is automatically selected.
Given a set of connection violations, the decision-support model identifies and enumerates all recovery options, including outbound flight delays, as well as connections for passengers and payload for each violation. There is a cost associated with each recovery option, and the model can support a complicated cost structure to reflect the reality. Thus, the model can help HCC controllers find the most cost-efficient way to recover connection violations.
The operations research team at Sabre Airline Solutions tested the model with one example having 19 connection violations. There were six outbound flights associated with these connections. The model considered the combination of all options for connection (e.g., allowing misconnection, normal transition and fast transition), as well as options for delaying outbound flights. It took less than one second for the model to find the optimal solution. Thus, this model can provide real-time decision support for HCC controllers in the recovery procedure.
The payload-management and turnaround-management models provide real-time decision support for HCCs. Airlines can use them to implement both turnaround and payload management for the HCC.
The two models can also be further enhanced by adding resource constraints to facilitate resource-management decisions.
For example, when tasks are sped up in turnaround management, an airline should consider the availability of ground resources. When more factors are considered in the decision-support model, better solutions can be obtained. Both models will be able to integrate with resource-management models, whereby all decisions will be made simultaneously and improve the decision quality.