[PDF] Revenue Management σε αεροπορικές εταιρείες Η

suggestions Keywords: Revenue management, airline companies, case study, linear programming model Revenue Management at American Airlines: A linear programming model solutions to complex and advanced decision problems

career in revenue management in 1984 with American Airlines He has served as Chairman of the Yield Management Study Group of the Airline Group of the

[PDF] REVENUE MANAGEMENT: DEVELOPMENT OF A CASE STUDY

study reports an analysis of the current status of how revenue management is taught in American Airlines reported $1 4 billion in revenue over a three-

[PDF] Yield Management at American Airlines

American Airlines estimates the quantifiable benefit at $1 4 bil- problem solutions are then combined to American Airlines yield management sys- one case and less valuable in the other By pleted long-term benefit studies, we antici-

[PDF] Yield Management - CORE

Albert V Casey, former President of American Airlines, Inc The following are beginning of modern day Revenue Management American's Solution was called "Super Saver Fares," which were Therefore, they began to study the behavioral In this case, we h o w that on average 15 people fly in class Y Using the

[PDF] Revenue Impacts of Airline Yield Management - CORE

17 jan 1992 · Only a very limited number of studies have been done on "American Airlines' yield management controllers are responsible for 38 million from the upper bound analysis and the No Control case will provide a useful tool to

[PDF] Revenue Management σε αεροπορικές εταιρείες Η - IKEE / AUTh

Bankov 2

Abstract

This research deals with revenue management. It aims to study the practice of revenue management, as it presents its definition, the historical background and the parameters of its application to the airline industry. The second part of the research presents a case study, using a linear programming model and having as an example airline company, American Airlines. The data for this research has been sourced from existing bibliography, while the linear programming model has been compiled using real data, such as aircraft capacity and the main hubs of the company. Finally, useful conclusions are presented as well as further research suggestions.

Keywords:

Revenue management, airline companies, case study, linear programming model

Bankov 3

Bankov 4

Abstract ........................................................................................................................................... 2

ȆİȡȓȜȘȥȘ ........................................................................................................................................ 3

Introduction ..................................................................................................................................... 5

Revenue Management Definition ................................................................................................... 7

Conditions for Applying Revenue Management ........................................................................ 7

Demand-Management Decisions addressed by Revenue Management ..................................... 9

Innovation in Revenue Management ........................................................................................ 11

Revenue Management History.................................................................................................. 13

The First Complete Revenue Management System: DINAMO ............................................... 14

The aviation costs breakdown................................................................................................... 15

Break-Even Load Factors.......................................................................................................... 16

Seat Configurations................................................................................................................... 16

Overbooking.............................................................................................................................. 17

Passenger Load Factor .............................................................................................................. 18

Linear Programming Model.......................................................................................................... 19

Revenue Management at American Airlines: A linear programming model ........................... 23

Model Outline ........................................................................................................................... 23

Demand and Capacity Constraints ............................................................................................ 30

Analyzing the Results Overbooking ...................................................................................... 33

Conclusions Further Research: .............................................................................................. 40

Appendix ....................................................................................................................................... 41

Bibliography.................................................................................................................................. 56

Bankov 5

Introduction

As an important chapter in Management Science, Revenue Management has been studied extensively over the last 50 years. Stemming from the deregulation of the airline industry in the United States and the entrance of fierce competition for aerial routes, this strategic decision- making process has been dominating the airline industry and has spread to other fields as well. Examples include the hospitality industry, with large chains of hotels, car rental businesses, the majority of which now operates on a global scale, as well as the food & beverage industry. As it is evident, yield management an early synonym for this practice all of the above industries

share the offering of services as their main income. The fact is that for these particular services the

companies operate on a relatively stable capacity (Kimes, 2005); their financial data presents a pattern of high fixed costs while variable costs remain low (Campbell, 2012); the product sold presents what is called a perishable inventory it can not be resold after production (Campbell,

2012); the industries also show a variability of demand in time a seasonality (Huefner & Largay,

2008); lastly, the ability to forecast demand, either based on historical data or mathematical

research is a characteristic of these industries (Huefner & Largay, 2008, Campbell, 2012). The current research provides an outlook on revenue management for the airline industry. After a bibliographical overview of the various decision-making variables, the history and definition of revenue management are studied. The incurred costs in the industry are broken down, to provide a clear picture of this vast industry which operates with low profit margins. The basic metrics are being defined, along with their importance as Key Performance Indicators for the operating businesses. The second part of this research provides an approach to the revenue management process by

Airlines. This

modelling and problem-solving procedure was seen in the work of Anderson et. al. (2012). In the current research, the model has been expanded to include significantly larger amounts of data. The statistical information from the United States Department of Transportation, Bureau of Transportation Statistics.

Bankov 6

A table of 60 flight products has been compiled, along with their fares, demand constraints as well as forecasted demand restrictions. The model has then been translated into SIMPLEX and with the usage of LINDO software, the task of maximizing the potential income for the flight products has been completed. The results also shade a light on overbooking. Overbooking is the practice of selling more seats than those physically available or overselling the inventory. It is a common practice throughout all airline companies. For American Airlines, overbooking is a standard part of their revenue management systems; the present study takes a look at overbooking by a single ticket for each flight product and not surprisingly, this practice is beneficial to the company. It is expected that the company has fine-tuned its inventory management, including overbooking practices, to avoid both empty, unsold seats, while simultaneously keeping the no-show income and overbooked passengers in their flights, minimizing any compensation. Finally, further study is suggested on the overbooking particularities as well as company policies regarding specific compensations.

Bankov 7

Revenue Management Definition

The practice of managing revenue streams, also called Revenue management is defined as the continuous usage of information systems while combining pricing strategies so that the right

capacity to the right customer at the right price and at the right time is allocated (Kimes and Wirtz,

2003).

the company, while delivering the outmost value to the customer using the service. In the aircraft industry, where revenue management (often referred to as Yield Management) is used predominantly, it can raise revenue gains up to 5% - profitability (Talluri and Van Ryzin, 2005).

Conditions for Applying Revenue Management

In the airline sector, what begun as Yield Management, has now expanded to a variety of industries: namely hospitality, retail and major service industries are applying revenue management models and techniques to maximize profits. Huefner and Largay (2008) quoting Kimes (2005) identify the basic assumptions that act as characteristics for an industry, or a company in particular, to apply Revenue Management successfully: Fixed Capacity: The company cannot easily change its capacity levels in order to cope with varying levels of demand or they can, but, more often than not, at a high cost. This condition is also seen as the less restrictive one for companies, as well as neither the most distinctive, since virtually almost any company has a fixed capacity of operations and service. High Fixed Costs with Low Variable Costs: The price structure of low variable and high fixed costs is a driving force behind adopting Revenue Management since any additional discounted sales would offer significant revenue. Campbell (2012) considers the example of the airline flight, where the costs of the aircraft, its crew and ground services would remain virtually unaffected with the addition of an additional passenger. The only addition would be mainly in fuel and on-board

Bankov 8

amenities. Therefore, selling more seats, which would otherwise remain empty is a revenue boost, at least in the short-term. While this requirement and characteristic may not apply to all companies and sectors, the latter would have to apply revenue management taking greater care in turning the increase of revenues to increase of actual profits. Perishable Service Capacity: Also seen as perishable inventory in recent bibliography, this is the characteristic that applies the service industries primarily. As Campbell (2012) states, the ability of a company to sell the services of its capacity is short lived and as such, unused capacity cannot be considered as inventory for future use or sale. Examples of perishable inventory and lost revenue can be a flight with empty seats on departure, a hotel night with vacant rooms, unfilled timeslots in restaurants, or golf courses and rental cars that remain available for the day. While product industries can generally hold back their unsold products to inventory, some physical products can also be considered as perishable and non-resellable, such as newspapers or dairy products with a short lifespan. Uncertain or Varying With Time Demand Patterns: Whenever demand is constant over time, it is met much easier with capacity by the company. However, as Campbell (2012) points out, agreeing with Huefner & Largay (2008), constant demand does not constitute a usual condition for revenue management applications, and demand varies across time. A typical example would include a hotel which sees high occupancy rates throughout the weekend, in contrast to weekdays. Airlines also see that their early morning and late afternoon flights on workdays are mostly sold-out by business travelers on single-day business trips, but midday and weekend flights remain mostly undersold. Ability to Forecast Demand: Having the ability to predict in advance the variable demand patterns, companies are able then to manage demand deploying variable pricing based on the ability to segment their customer base into groups, in order to determine which group would then receive any specific price offers and reductions. Since the companies want to avoid offering lower prices to customers willing to pay more for the same product, they place terms and restrictions to the lower price classes, thus segmenting the customers. For instance, restaurants offer early dinner

Bankov 9

discounts and happy hours to retirees and similar customers with time-flexibility during the day. Hotels and airlines use their advance purchase, non-refundable rates to segment leisure from business travelers, since the latter ones have time constraints, as well as the willingness to pay more for a specific flight which suits them. Albeit this is an imperfect segmentation these often

especially for hospitality and airlines is a characteristic of businesses that utilize revenue

management (Huefner & Largay, 2008), while the goal would be to attract new potential customers to the business rather than allocating different groups and rate classes to existing customers, who would otherwise be able to pay full price for the product (Campbell, 2012). For early applications of revenue management, forecasting was facilitated by reservations, or advance selling of tickets. Out of those five conditions, three are more generic and less restrictive. Having a fixed capacity, the ability to forecast, even inaccurately demand and the variability and uncertainty of demand can be found, at varying parameters, in any business. However, a cost structure dominated by high

fixed costs allows for a company to ignore the marginal costs incurred in their pursuit for additional

revenues and as Campbell (2012) states, added revenue would approximate to added profits in this case. The last restrictive parameter is that of perishable, non-resalable inventory. This condition eliminates the usage of an inventory policy that companies in other sectors have and sets off the usage of revenue management as a specific strategy in itself. Demand-Management Decisions addressed by Revenue Management As Zatta (2016) agrees with Talluri and Van Ryzin (2004) that revenue management is directed towards three fundamental demand-management decisions: Structural decisions: which will be the differentiating and segmenting mechanisms and characteristics included, if any; which selling format is to be used, negotiations, fixed prices, or auctions; which terms and conditions are to be offered, such as bulk discounts, cancellation and refunds or no-show charges; how are products to be bundled, and others. Pricing decisions: how are posted, individual or reserve prices (in case of auctions) set; how is pricing decided along a variety of different categories of products; over the passing of time, how

Bankov 10

will the price be defined; how is discount pricing (markdown) set over the product life cycle, if so,

among others. Quantity decisions: whether an offer to buy would be accepted or rejected; how would output and capacity be allocated optimally to different segments, product offer and channels; when a product is to be withheld from the market and sold at a later point in time, and other decisions. Whichever of the above decisions comes first in terms of importance, this will depend on the wider context for the business itself. What also varies is the timescale these decisions are taken. For instance, structural decisions regarding the mechanisms for segmentation selected, are taken infrequently, since companies commit to their products for the long-term. Commitments can also be made in pricing and quantity decisions as well, since firms may have to advertise pricing in advance, or deploy their capacity in advance: this would hinder their ability of adjusting prices or quantities on an operational level. Companies also bear the burden of capacity and inventory reallocation. In particular, as Zatta (2016) and Talluri and Van Ryzin (2004) agree, the usage of capacity controls in airlines derives that a company is selling (various ticket types being sold at a plethora of times and bearing different terms and conditions) are all supplied using the same,

consistent aircraft seat inventory. Consequently, airlines are given enormous flexibility in terms of

quantity, and quantity control is a predominant strategy in the industry. Whereas, retail businesses would, often, commit to the quantities they would be selling (with the initial stocking decisions), while they can present more flexibility in fluctuating and adjusting pricing over time. However, as both researches indicate, the ability to price in a tactical manner depends on how costly prices changes can be which can vary depending on the distribution channel, such as online or catalog. Whether a company would be using price-based or quantity-based Revenue Management, could vary even across companies within a specific industry. For instance, while the majority of airline companies are tied to presenting fixed prices while also allocating their capacity in a tactical manner, low-cost carriers regularly have their pricing as their foremost tactical variable. The businesses involved can find innovative ways in which to increase their power to make quantity or price resource decisions. An example Talluri and Van Ryzin (2004) point out is for retailers, who can hold back an amount of their stock in a centralized storage unit and proceed to make a mid-season refill decision, rather than precommitting their entire stock to the final retail

Bankov 11

points. Within the airline industry, some companies have experimented with movable seat arrangements which allow them to repartition the business and economy allocation, on a short- term basis. Other major airlines have practiced demand-driven dispatch, noted as D3, in which aircraft of different sizes are dynamically assigned to each flight departure, responding to demand fluctuations and thus not being allocated and committed in advance. As regards pricing, the usage of online channels provides the companies with more flexibility. All the aforementioned innovations increase the opportunity for quantity and price-based revenue management.

Innovation in Revenue Management

The practice of revenue management can be identified as a very old concept. Throughout the history of trade, every seller has, at one point or another, have come across revenue management type questions. What would the best price to ask? Which offers should they accept? When would a discount be available? And, when would be the time to stop and try selling at a later point in time, or in a different market. These problems that revenue management addresses are as old as the practice of business itself. In the theoretical level, the problems that revenue management addresses are not new either. The forces of supply and demand as well as that of price formation that comes as a result are still the foundation of our knowledge regarding market economics. And the concept of the profit- mechanisms by which market equilibria can be reached. Modern economic theories address a large number of advanced and subtle demand-management decision making: bundling, segmentation, non-linear pricing, price optimization in the presence of varying information and data between buyers and sellers, are to name a few. The innovation that revenue management brings, is not found in the demand-management decision themselves; it is found in how these decisions are being made. The actual innovative character of revenue management regards the method of the decision-making process, which is now a highly sophisticated, technologically evolved, detailed, precise and an increasingly intensive operational approach to reaching these demand-management type decisions.

Bankov 12

The revenue management approach is driven by two key factors. Firstly, advances in the sciences of economics, statistics and operational research now make the modelling of economic conditions and demand possible, as well as to quantify any uncertainties faced by the decision makers, the estimation and forecasting of market responses, the computing of the optimal or near-optimal solutions to complex and advanced decision problems. Secondly, advances in information technology can now enable the automation of transactions, obtaining and the storage of vast data to such an extent that the advent of Big Data is discussed quickly execute complex algorithms and implement and control detailed decisions regarding demand management [Zatta (2016), Talluri and Van Ryzin (2004)]. Therefore, combining science with technology in demand- management remains the typical feature in yield management. Both factors are relatively new as well: most of the science used forecasting and demand models, specific optimization algorithms are relatively new, less than 100 years old. While the information technology online databases, data collection via internet, software are less than 50 years old. These advances have transformed the unimaginable concept of accurate modelling of real world phenomena into revenue based decision making, thus operationalizing the science. The combination of these two key elements has brought acute benefits and consequences to demand management. The first consequence is the ability to manage demand on a large scale and complexity, to levels that were considered as unreachable, only a few decades ago. Instances of this include airliners

with hundreds of flights per day, servicing hundreds of airports and destination pairs, each of which

is sold at multiple prizes. Similarly, as Zatta (2016) notices, large retailers operate more than a few

thousand product codes which are sold in hundreds of stores and via online channels, updating pricing on a daily basis at least. As Van Ryzin (2004) argues, the scale and complexity are so vast

in instances such as these, that decision-making is beyond the ability of our natural capacity. If it

were not automated, it is required that it be aggregated at a such level and thus simplified, where significant opportunities for incremental gains in specific locations, channels and products in time would be lost.

Bankov 13

Revenue Management History

The Civil Aeronautics Board (CAB) of the USA had been regulating, since 1938, all air transport

routes as a public utility service, on a federal level. Acting as the regulatory authority for air travel

between states, the CAB would define the routes service, as well as their specific schedules and fares. Part of the CAB mandate was to ensure a reasonable profit for the airlines involved in passenger and cargo transport, as well as to regulate new entries in the market. This depression- era approach was considered ineffective and costly, especially with the advent of jet airplanes in by the states themselves, some states such as California and Texas were not enforcing any regulations. Hearings presided by the US Senate Judiciary Committee discovered that in 1975 the San Francisco Los Angeles fare would be half than that of the similar and lawfully monitored Boston Washington itinerary. At the same time, an intra-Texas domestic airline advertised that they would help farmers cross the state by flying, with less expense than driving, including bringing their own farming equipment on board (Breyer, 2011). The Airline Deregulation Act of 1978 introduced a free market for air transport in the USA, removing federal government control and influence over routes, fares and market entry for new competitors. This increased competition among air carriers sharply it also meant that profit is not guaranteed to the competitors. In 1974, 207.5 million passengers were flown, while in 2010 the figure was 721.1 million. For the United States, out of the twenty-two legacy carriers, airline companies operating before the Deregulation Act of 1978, only five of them remain operational in Airline revenue per passenger mile has declined from an inflation-adjusted 33.3 cents in 1974, to 13 cents in the first half of 2010 American Airlines was the first airline to implement a revenue management system in 1985. The advent of People Express and other low-cost airlines posed a threat to the company; however, as Campbell (2015) mentions, American could not match the lower prices, since that would be a major factor for loss of revenue. The company focused instead on ways and processes to target reduced fares to flights, times and customers, aiming to have a heavy impact on its competitors, while keeping its usual price structure in other circumstances. While customers welcomed the

Bankov 14

pricing competition, they preferred to fly with the proven companies, since their reputation was already high American managed to outcompete the low-cost carriers of that time. their leisure and business clients separately, thus segmenting their customer base, it was not until management became quite clearer; selling enough lower-priced seats to leisure passengers would cover the fixed costs while the higher prices to their business travelers should maximize their revenue (Tyrell, 2017).

As Tyrell (2017) explains it, in order to have a measurement of the success of this pricing strategy,

nue per passenger or passenger percentage of seats filled metric. Thus, the practice of yield management was created. Nowadays, airlines use the RASM metric which translates to Revenue per Available Seat Miles, amongst others, such as the Load

Factor percentages and Yield in cents.

The First Complete Revenue Management System: DINAMO complete methodological approach to yield management. The development of its highly sophisticated and complex platform was the only solution to a revenue management decision making problem which is best described as a nonlinear, stochastic mixed-integer. It also requires data that includes passenger demand, ticket cancellations and other passenger behavior, all subject to periodic changes within the day. (Smith et. al., 1992). These decision variables, which Smith estimates that amount to approximately 250 million, required the development of operational research models. The American Airlines Decision Technologies (AADT) sector of the company, decided to reduce the original large problem into a number of smaller ones, all more easily manageable and realistically solvable, which also model the real-world conditions and situations. According to the same authors, the company added more than 1.4 billion dollars to their net income for the period 1989 1992.

Bankov 15

The effects of this new system were the main reason why PeopleExpress, the major low-cost competitor of American Airlines was sold off to Continental Airlines, after mounting losses over consecutive years. The CEO of PeopleExpress states: profitable company from 1981 to 1985, and then we tipped right over into losing $50 million a month. widespread Yield Management in every one of our markets. We had been profitable from the day we started until American came at us with Ultimate Super Savers. That was the end of our run because they were able to under-price us at will and surreptitiously. Spoilage is defined as loss revenue, or empty seats on a flight. On its first years of operation, DINAMO reduced spoilage to only 3%, while Yield Management Analyst production increased by 30% (Donovan, 2005).

The aviation costs breakdown

The Department of Transportation of the US Government has broken down the costs of airline operations, in the following order: Flying Operations: These include all costs incurred with the operating of the airplane. wages 37% Transport Related: The second largest category includes any outsourced regional capacity provided and the cost of in-flight sales 17% Aircraft and Traffic Service: Arguably the most extensive costs, since they include ground handling for passengers, cargo and aircraft and includes salaries of baggage handlers, dispatchers and all ground crew 14% Maintenance: The costs for parts and labor are high, since airline operations are safety intensive 10% Passenger Service: The in-flight service provided, especially by legacy aircraft carriers can be costly, since it includes complimentary meals and the flight attend 6% Sales and Marketing: A broader category which includes advertising, reservation and travel 6%

Bankov 16

Administrative costs: These amount to 5%

Depreciation / Amortization: A category which includes technological equipment and plants 5% The most common costs, which apply to almost all the above categories are the labor costs. When considered in their entirety, human costs, such as wages, account for 25% of the operational expenses, and 75% of the controlled costs. Only recently (since 2000), has fuel overtaken labor costs to become their most important cost from 25 to 30 percent of the total expenses, while transport-related expenses come 3rd, at 17 percent. It has to be noted, however, that transport-related costs have risen substantially over the last few years, leading many major

airliners in the position to outsource a big percentage of their schedules to smaller, regional carriers

in a move to allocate existing supply and costs more closely to demand (ATA, 2008).

Break-Even Load Factors

In order to become more specific, every airline and thus every flight has a specific break-even load

factor. This would be the percentage of the seats, that the airline has in active service and that must

sell at a specific price level, also called yield, to encompass its associated costs. As one might expect, revenue and costs differ from company to company; therefore, higher costs would in turn increase the break-even factor, while selling at higher fares would have an adverse influence. Unfortunately, due to rising fuel prices and lower average fares, the break-even load factor remains above 80 percent (ATA, 2008). As such, airlines operate vert close to their load factor and the sale of a few more seats on each flight can actually become the steady distinction between operational profit or loss.

Seat Configurations

The addition of inventory to an aircraft is an option to increase revenue at a relatively minor

marginal cost. It has to be noted, however, that the optimal seat configuration of an aircraft actually

Bankov 17

while keeping prices the lowest they can be, while increasing seat inventory is an optimal strategy for leisure travelers and low-cost travelers. While maintaining a premium economy diversification of seats and opting for less dense seat layout is a strategy which targets business travelers. The key for most airlines is striking the right equilibrium, since most serve a broader blend of business and leisure travelers.

Overbooking

In order to maximize incomes throughout the network while keeping as many customers as possible, an airline may choose to overbook flights; this means that it would book an abundance of passengers than they have availability for, on a given flight. This is partly selected to balance After analyzing the historic demands for a flight, economics and human behavior,

this practice compensates for the fact that some customers, particularly business travelers buy full-

that they would have to book different flights or cancel their trips, frequently without noticing the

airline company. They might as well be delayed due to increased numbers or long security queues, or even reserve seats on more than one flight. It is in the best interest of both customers and companies to have all the seats sold, when the reservations are being made. Since, as stated, an airline seat is a product that is perishable and

when a customer does not attend to their flight, that specific seat cannot be transferred as inventory

for future use, airline productivity is undermined, which contributes to higher fares and limited service. As a consequence, some airlines may decide to overbook flights this is not done however in a random and aimless fashion. In practice, companies use vast historical data of flights and determine the number of seats authorized to be sold. The ultimate goal is to equalize the amount of seats overbooked to that of no-show passengers. More often than not, this practice is functional and effective. On some occasions, when a surplus

of people show up for a flight, than the seats of the aircraft, airlines provide incentives and benefits

to passengers, so that the latter renounce their seats. Travel vouchers, overnight accommodation and other forms of compensation may act as incentives for passengers being re-booked on another

Bankov 18

flight. When not enough passengers volunteer for these, airlines will have to bump them involuntarily, offering law-regulated compensations, on a case-by-case basis.

Passenger Load Factor

Passenger Load Factor (PLF) is a fundamental metric for the airline industry, since it measures

be used to compare airlines of different sizes and capacities, since this metric measures the relative

success of the companies to fill up their aircraft seats, or inventory. PLF can be calculated by diving the Revenue per kilometers for passenger (or Revenue Passenger Kilometers) by the Available seats per kilometers for the flights (ASK). This metric is considered important (Jadhav,

2019) because the companies are looking to have it maximized, while its trends help senior

management to take decisions regarding pricing, capacity as well as frequency of flights (Jadhav,

2019) and scheduling. The upwards or downwards trends, provide a clear answer to whether or

not an aircraft is flying empty and if any corrective actions are needed. It can also mean that a f number of kilometers flown (Jadhav,

2019).

Bankov 19

Linear Programming Model

While the development a revenue management system is an expensive, labor intensive and lengthy procedure or a big company it cannot be stressed enough that the potential payoffs are substantial. As we have already discussed operational research has been in use since the 1960s from airline companies. The most popular applications of operations research in these businesses addressed issues such as aircraft and crew schedule planning, Including schedule design which defines which markets are to be served with what frequency and how are flights to be scheduled in order to meet these demands, Fleet assignment which could help specify what size of airplane is to be assigned

for each flight, as well as aircraft maintenance routing which determines how is a flight routed to

and from its technical center ensuring satisfaction of maintenance requirement as well as minimizing costs incurred for the company And Crew Scheduling which would select the cruise to be assigned for each flight thus minimizing crew costs (Barnhard et. al., 2003). As Barnhart et. al. state, applying operational research for solutions of problems such as fleet assignment, has resulted in the savings of at least a hundred million dollars for Delta Airlines, while American Airlines has been reporting a steady Improvement in operating margins. It is also concluded that Conducting research on such issues has now provided sophisticated techniques for the solution of general linear programs. One such instance was the node consolidation idea, which was introduced in 1995. It managed to reduce the formulation size for the problems of one major airlines by more than 40%. This model has now been generalized and encapsulated into commercial soldiers which in turn allows for solving more efficiently, optimization questions at a larger scale. As a strategic decision, airlines choose to overbook reservations, above capacity limits, thus overbooking their flights for more than two decades (Barnhart et. al., 2003). This is a strategy in order to reduce potential loss of income and revenue stemming from passengers who do not show up. With the incorporation of Revenue management IT systems, overbooking is now a part of the seat inventory control functions of the latter. The question which is posed, in terms of overbooking for these Revenue Management Systems is to identify which is the upper limit of acceptable reservations for any flight departing, while balancing the risks which come, as well as costs in

Bankov 20

denied boardings, offsetting any probable losses in revenue that come from seats that were no sold, or are characterized as spoiled. Using statistics, the early models for overbooking usually show the no-show percentages as Gaussian random variables. Their objective is to find the upper limit of authorized number of bookings which would maintain denied boardings below an acceptable level, as specified by the companies. A continuation of the statistical overbooking approach is the post based overbooking model; this model clearly states the real cost incurred by denied boardings and the actual costs incurred when an aircraft is flying with empty seats (spoilage).

As Barnhart et. al. (2003) notice, this latter cost-based model for overbooking is the latest practice

for the majority of airline companies. It has been deemed as effective and accurate. It should be noted however that this approach represents a static formulation of the overbooking problem; and the dynamics of passenger bookings, cancellations and no-shows are not explicitly accounted for in order to determine an acceptable overbooking level (Barnhart et. al., 2003). While in practice the revenue management topic has been covered extensively in operational research literature and this has led to dynamic programming formulations, such as a two-class, joint overbooking and fare class mix problem, only a small number of airlines are now using complex dynamic programming models due to the difficulties in having accuracy and adequacy in inputs in terms of reservation and cancellation percentages by the defined time before the flight has departed. It should be noted that the financial leverage for booking above the limits in the industry is quite substantial. For instance, in the USA, intra-country no-show percentages are in average, 15 to 25% of the last pre- departure reservations. Bearing in mind that for the majority of businesses in the industry, the aim

is to attain a 5% margin of operations, this loss of 15 to 25% of possible revenues in popular flights,

which are fully booked (occurring without overbooking) can represent a major cost factor on their profits. Therefore, bearing an effective yield management system, managing overbooking has proved that it produces the same revenue that comes from optimizing allocation in the best fares possible. As Barnhart et. al. (2003) mention, another major technique of Airline Revenue Management revenue maximizing mixture of seats to be made available for selling per booking fare or class on each precise flight departing. As such, every Airline Revenue Management System has been made in order to optimize this fare mix as their main goal.

Bankov 21

These Revenue Management Systems can forecast expected demand separately for each future flight departure, through the application of statistical models contrasted with historic reservation points for the similar fare class on previous flights of the same itinerary. These are then used as input information to the asset allocation optimization model which can determine booking limits; these booking limitations are being used separately to every booking class of every flight under examination. Most of these systems presently help develop structural inventory allocations relying on the usage of serial nesting of reservation classes, seen below. Figure 1: Nested Booking Limits (Barnhart et. al. 2003). The serial nesting system works as such: Instead of allocating seats to partitioned classes within an aircraft, seats are actually keptmore expensive fare classes while nested booking limits are used to the fares that are the lowest. Every seat which is free in this larger, common inventory can sold at the primary, most expensive booking category in the, rather rare, event that the whole airplane is full of passengers demanding the most expensive flight product for sale (i.e. all passengers of an aircraft book first-class tickets). This happens so that the airline can ensure it would not deny a booking at the upper fare, when there is inventory in the aircraft to be used for the flight itse

Bankov 22

fundamental rule which protects the full fare seats was extended by Belobaba (1987, 1989) to include multiple fare classes. The latter researched the first PhD thesis on pricing in the industry and yield management. These models which rely on decision rules that are heuristic for these nests of classes and have now been regarded as frequent models for controlling aircraft inventory in Airline Revenue management decision-making systems. Nowadays, managing network revenue, also called origin-destination control, represents a further advancement, going beyond the fare class mix capabilities which most third generation Revenue Management Systems control. It is under development and being tested and used by the largest

airlines globally. As it is named, origin-destination control can provide the airline with the ability

to control its seats with the real income flight network, and not merely following the asked for fare class on an isolated flight. Since the world's largest airlines operate on a complicated spoke and hub network, or a series of these, an important percentage of their passengers purchase multiple flight legs through a hub connection. American Airlines was the first in implementing an origin-demand control strategy The company used the technique that became known as virtual nesting. This is a process of assignmentquotesdbs_dbs20.pdfusesText_26

[PDF] [PDF] Revenue Management σε αεροπορικές εταιρείες Η - IKEE / AUTh

[PDF] Revenue Management and Pricing: Case Studies - Veritec Solutions