Designed For The Future
Optimized Augmentation: A New Milestone In Crew Optimization
Enhanced and improved crew science is essential to efficient airline operations. With a greatly advanced pairing optimizer, the potential for carriers to optimize crew scheduling throughout various aircraft fleets can be game changing.
For airlines, planning the schedule and operations is a multistep decision-making process, with various time horizons. Typically, such steps revolve around schedule development, fleet assignment, aircraft maintenance routing and crew scheduling.
Flight-crew expenses and fuel costs represent a significant portion of an airline’s direct operating costs. Hence, the primary objective of the crew-scheduling problem is to find and match up a minimum cost assignment of flight crews to a given flight schedule, while at the same time complying with various government regulations and union work rules that are designed to ensure flight safety and maintain crewmembers’ quality of life.
Next-Generation Crew-Pairing Solution
Sabre has been a thought leader in crew-pairing optimization since the early 1990s. The Sabre Long-Haul Pairing Optimizer has evolved from using an all-inclusive MIP approach to using state-of-the-art column-generation techniques with a primal-dual-based solution approach. The most recent innovation, one-pairing optimizer, uses dynamic augmentation with column generation, setting this solution apart from its competition.
Because of the extreme complexity involved, the process has to be performed in two fundamental steps: 1) Optimized anonymous crew pairings must be generated, following which 2) they can then be assigned to crewmembers to cover the given flight schedule.
Some basic definitions: A pairing is a tour of duties that begins and ends at a crew base. The duties in a pairing are separated by rest periods. A duty consists of a sequence of flights between check-in and check-out, which are separated by connection times.
Analyzing The Problem
Traditionally, for flight-deck crewmembers, the crew-pairing-optimization problem is solved separately for each aircraft fleet, since flight-deck crewmembers are not always trained across fleets.
The crew-pairing-optimization problem for short-haul fleets, characterized by their short flight times but dense flying schedule, is typically solved using commercial crew-pairing-optimization systems, assuming a fixed cockpit-crew requirement on each flight (in other words, one captain and one first officer, referenced henceforth as “basic crew”).
When it comes to long-haul flights, government and union regulations usually do not allow operating with only a basic crew. Some duties may require extra crew, because of their excessive duration or the unacclimatization of crewmembers, to ensure reduced crew fatigue and safe flight operations.
A demonstrated industry breakthrough, Sabre Long-Haul Pairing Optimizer has gone above and beyond current industry standards, with capability to generate optimized pairings by solving dynamic augmentation, duty-selection and pairing selection decisions in the same model, removing suboptimal guesswork. In its approach, Sabre factored deadheading, downranking, crew-augmentation and crew-splitting decisions into a single operations-research model.
By Sabre estimates, cost savings for the airlines that have implemented the Sabre Long-Haul Pairing Optimizer ranged from 5 percent to 11 percent over their existing setup.
Sabre Long-Haul Pairing Optimizer has traditionally not dealt with extremely complex large-scale short-haul networks for augmentation-related decisions. However, recent or pending changes to flight, duty and rest regulations for flight and cabin crewmembers, especially in short-haul operations, are posing challenges to airlines worldwide.
FAR Part 117 Rules
Airlines across the United States have recently been mandated to use Federal Aviation Regulations (FAR) Part 117 rules for their passenger-operation flights, geared to address fatigue dealt with by crew.
The new FAR part 117 rules make many changes to flight, duty and rest duration limits. They place more stringent limits on duty periods that begin or span known periods of circadian lows, or time periods during which the human body has its most likely tendencies to require sleep.
The final release of FAR Part 117 addressed some of the initial comments from the industry, such as the general concern about a stated assumption of 25 percent cost savings that have been predicted to result from long-term scheduling optimization in the U.S. Federal Aviation Administration’s (FAA) regulatory impact analysis (RIA).
A subsequent FAA analysis, which applied the principles of a commercial crew-pairing optimization system using basic crew, found that airlines don’t need to hire additional crewmembers to comply with the new rules (because flight crewmembers currently used in reserve enable carriers to conduct operations under the rules with no additional crew).
The analysis also estimated a nominal cost of US$390 million as the cost of compliance with the rules over a 10-year period, while the nominal benefits ranged from US$376 million to US$716 million.
Airlines in general have refrained from publicly characterizing or tallying the impact of FAR Part 117 rule compliance on workforce issues and profitability. Nonetheless, there are scattered (and generally inconclusive) reports estimating 1 percent to 2 percent cost increases due to compliance factors.
Interestingly, a section of the FAR Part 117 rules that did not generate much analysis is the new classification of onboard rest facilities and in-flight rest requirements, which allow even shorthaul fleets to be potential candidates for augmenting crewmembers to extend duty periods while mitigating fatigue-related risks.
Among rest facility classifications, a Class 3 rest facility is described as a seat in an aircraft cabin or flight deck that reclines at least 40 degrees and provides leg and foot support.
Most business-class seats on shorthaul flights with or without marginal reconfiguration meet this requirement, hence opening up an interesting question for crew-planning personnel: If one augments a domestic short-haul U.S. duty with an extra pilot, or two extra pilots, are there any cost savings to be derived without compromising in the area of crew fatigue?
Can it compensate for any extra cost that FAR Part 117 compliance requires?
In a substantive effort to answer the first question, Sabre has developed a new one-pairing optimizer solution that includes the optimized dynamic augmentation capabilities along with downranking and crew-splitting features into our most sophisticated optimizer that uses stateof-the-art network reduction and column generation techniques to find provable optimal pairings in the largest of the airline problems.
Applying the new capabilities of Sabre’s new one-pairing optimizer, we conducted a case study with the help of a U.S.-based airline.
Significant Crew-Pairing Cost Savings
In a recent case study, one month’s optimized pairings were created for the participating airline’s Boeing 737 fleet. Baseline results, which did not use dynamic augmentation, were marginally better than the commercial solver in use at the airline. However, when using dynamic-augmentation capabilities with Class 3 rest facility allowed in FAR part 117 rules, onepairing optimizer produced impressive cost savings.
Case Study Specifics
The case study involved one month’s schedule for the participating airline’s biggest domestic short-haul fleet (which accounts for 60 percent of the carrier’s overall operations).
The scenario analyzed had about 900 operating flights to be covered per day with about 5,000 deadhead flight options per day, including all flights from other fleets and subsidiaries of the airline.
Elements applied included changing Sabre’s internal rules engine, historically named pairing legality check (PLC), to incorporate the FAR Part 117 rules, along with additional airline-specific restrictions and buffers (in this study, the airlinespecific rules were much more restrictive than FAR Part 117 rules).
Also applied was airline-provided costing to evaluate the pairing solution, which included accumulated credit-hour costs, nightly hotel costs, per-diem costs and deadhead costs, as well as the airline- provided pairing solution using basic crew generated by a commercial pairing-optimization system used at the participating airline.
An advanced commercial-pairing optimization system like the one used at the participating airline, as well as Sabre’s system, can generally solve pairing-optimization problems very close to optimality. To establish the competitiveness and validity of the one-pairing optimizer, a baseline solution was first generated using the same parameters with regard to costing, rules and manpower restrictions applied by the airline in its pairing solution with the basic crew.
Then, having established a comparable baseline solution with the one-pairing optimizer, two augmentation analyses were performed on the short-haul fleet. For the first optimized-augmentation analysis, the airline regulations allowed use of only the block-hour extensions and no flight-duty period extensions. In the second analysis, the flight-duty period was added based on augmentation usage. In both analyses, airline-provided buffers were applied. The rules used for both scenarios are more stringent than FAR Part 117 limits. The potential savings could be even greater if we used the FAR Part 117 directly.
For both analyses, the one-pairing optimizer was run without applying downranking (meaning that a higher-ranked crewmember, such as a captain, might take a first-officer, or lower-ranked, role) or crew-splitting concepts.
In the first analysis, even with heavy buffers built on block-hour extension rules and no flight-duty period extension allowed, the one-pairing optimizer was 0.7 percent better (US$3.89 million per year savings) compared to the airline-provided solution, and the one-pairing optimizer was 0.5 percent better (US$2.54 million per year savings) than Sabre’s baseline solution.
In addition, the one-pairing optimizer found that using only single extra crew is advantageous in comparison to double extra crew.
In the second analysis, the solution generated by one-pairing optimizer was 1.3 percent better (US$6.4 million per year savings) over the airline-provided solution, and results were 1.1 percent better (US$5.3 million per year savings) than Sabre’s baseline solution.
The Path Forward
Results of the case study reveal the potential savings to be realized using the concept of pairing optimization based on dynamic augmentation, while retaining all negotiated crew-union restrictions for U.S.-based airlines.
Not only are the potential savings huge, but application of this process can help make crewmembers more relaxed during operations and can, thereby, mitigate any cost increases due to compliance with FAR Part 117 rules.
From a pure research perspective, these results are extremely powerful, as flight-deck pairing optimization has remained static for many years.
For airlines, having the knowledge of potential increases in crew utilization, crew productivity and pairing-operating costs (deadhead, hotel, crew meal, per diem, etc.) may identify instant new pair- ing-construction strategies for immediate implementation, or at least provide information to serve as the basis for carriers to conduct meaningful discussions with appropriate entities to enable application of such a strategy.