The document discusses using a genetic algorithm to optimize airline revenue management. It proposes applying genetic algorithms to determine which ticket booking terminals should remain open to maximize profit. The algorithm would consider historical booking data, customer demand forecasts, and ticket prices to select the terminals where booking higher fare tickets would generate the most revenue. It provides an example case study demonstrating how the genetic algorithm evaluates multiple booking requests to choose the terminal that contributes most to overall profit. The results show the genetic algorithm approach can increase total revenue compared to a simple first-come, first-served booking system.
2. ABSTRACT
Basically revenue management is designed to increase the annual profit of any organization.
In this paper Revenue management of airline industry at the time of ticket booking is discussed
by use of genetic algorithm .A genetic algorithm (GA) is a method for solving both constrained
and unconstrained optimization problems based on a natural selection process that shows
similarities to biological evolution. In that as a result it is decided that to show unavailability of
the tickets at the terminal having lower cost of ticket value rather than to close the counter of
terminal , because closing of terminal leads to reduction of revenue in the case of other
destination flight tickets. To get these results genetic algorithm is applied step by step & to
manage the Revenue Management system.
3. Introduction
In the area of revenue management, the main aim is increase the revenue. To get this we must
understand what is revenue management and what are the steps to implement the revenue
management system.
What is Revenue Management?
The revenue management is also be known as yield management, but basis difference in these is
Revenue Management is more strategical and Yield Management is more tactical,Revenue
Management also allows us to apply this in a wide area.
“Revenue Management is the art and science of predicting real-time customer demand at the
micro market and optimizing the price and availability of products” [Cross, R.G. (1997)].
Revenue Management (RM) represents the technique to analyze or to optimize revenue of a
certain company by selling the right (quantity of) products at the right time, for the right price
and to the right customer.
The important steps to implement a Revenue Management system are:
1. Building a business case and analyzing – it helps to understand the level of investment and to
understand all aspects of the business to the company.
2. Product price estimation- Pricing has become an important part of the business because it is
important to understand how much products and services its customers buy as opposed to the
price the company is offering
3. Buying or building the system- buying RM system means external expert come to the
company to analyze the business and products. But some of the companies don’t allow external
experts to come in company due to their personal reasons, then instead of buying, they decide to
build the RM system but that has a disadvantage as mistakes can happen often as lack of
experienced personnel in the management
4. 4. Data storing and forecasting- the data of analysis is get stored continuously over the years.
And that data is get used for the future forecasting (to predict consequence in future).
But as in case of airline revenue management following steps must be covered to implement
RM-
1. Market segmentation
2. Price management
3. Demand forecasting
4. Availability
5. Reservation negotiation
Most of the RM systems uses the GA technique that is Genetic Algorithm.
Genetic Algorithm
A genetic algorithm (GA) is a method for solving both constrained and unconstrained
optimization problems based on a natural selection process that mimics (shows similar as )
biological evolution. . Genetic Algorithms(GAs)was invented by John Holland and developed this
idea in his book “Adaptation in natural and artificial systems”.
Genetic Algorithm is a powerful tool for solving search and optimization problems of various
industries. Genetic algorithms are based on the principle of genetics and evolution.
Some terminologies used in GA as follows-
1. Chromosome – it is raw genetic information that GA deals with (sequence of gene).
2. Individuals – it is a single solution to each of the chromosome.
3. Gene – it is basic set of instructions to build GA.
4. Fitness - The fitness of an individual in a genetic algorithm is the value of an objective
function for its genetic information.
5. Population - A population is a collection of individuals.
5. The steps to be followed in GA are as follows-
1. Selection of a population.
2. Assigning the fitness value.
3. Breeding – this process consist of three steps first is a selection of parent, second is a
crossing the parent to create new individuals (crossover) and third is replacing the old
individuals with new ones.
4. Termination .
The population size and the fitness value changes form application to application.
6. Literature review
1. A. Tolk, S. Y. Diallo, I. O. Ryzhov, L. Yilmaz, S. Buckley, and J. A. Miller, eds.
GENETIC ALGORITHMS FOR CALIBRATING AIRLINE REVENUE
MANAGEMENT SIMULATIONS
The basic aim of a revenue management system is to maximize the revenue
i.e annual income of any company or organization. In this paper the Revenue
management(RM) theory is done by simulation modelling by using Genetic Algorithm
(GA).
Simulation modeling allows RM researchers and practitioners to analyze the
potential consequences (effects in future) of new strategies. Simulations can be used
to evaluate new forecast methods and optimization algorithms.In this paper attention is
towards agent based modelling. The simulation system used in this paper, REMATE
(Revenue Management Training for Experts), has been developed at Deutsche
Lufthansa in cooperation with several German universities since 2009.
REMATE models airlines and customers as agents interacting in dynamic markets
(Cleophas 2012a). The simulation employs a stochastic, discrete-event-based paradigm
and implements state-of-the-art algorithms for Revenue management (RM) forecasting
and optimization to model airline Revenue management (RM).
In genetic algorithm approach population is get selected each individual
gives a different solution as genetic codes we can also called it as fitness value, then
individuals having best fitness value gets selected and are cross-bred to generate new
instances. followed by mutation converges to the required or optimum solution.
Application of GA is done here by two ways.
1. Calibrating analyst influence.
2. Calibrating demand.
The results found from above are as follows
The system including GA set analyst influences systematically out performed system
lacking influences.
Demand calibration using GA gives an improved model fit (best result) and proved to
be significantly consumes less time than manual calibration.
As a result, Genetic algorithm turs out to be versatile tool for calibrating agent-based
simulations.
7. 2. A Genetic Simulator for Airline Yield (revenue) Management Vol. 2 Issue 9,
September – 2013
The main concept of yield management is to serve or to provide the right
service to the right customer at the right time for the right price. Mainly this method is
used for revenue management. In this paper author mainly introduced yield
management which includes data collection, segmentation, forecasting, optimization,
and dynamic re-evaluation. Author has also given some of examples in this papers
which proves the effect of Revenue Management and decision-support computer
system This system has helped to monitor all historical data on the company flights as
well as has made it possible to perform flights forecasts up to a one year period with
high precision.
To maximize the revenue of airline, using Genetic Algorithm an optimized
flight booking and transportation terminal open/close decision system has been
presented. In this system, the particular booking terminal’s historical booking data is
observed and recorded. At the same time its frequency is generated with linguistic
variable and deviation of booking is interpreted. Using the observed data and genetic
algorithm, the terminal open/close decision system is optimized. While doing this also
needstobe consider some parameters like the effect of time-dependent demand, ticket
cancellations and overbooking policies.
Genetic algorithm done by following steps-
1. Initial population
2. Parent selection (by fitness value)
3. Reproduction (crossover)
4. Mutation
5. Termination.
At the point of termination the algorithm will usually return the best individual
according to its fitness function value. but termination is done by either when objective is reached
or reached to external conditions like when fixed number of generations reached & maximum time
allowed reached.
As the results obtained although can prove to be useful for airlines industry but still
there are a number of things that can be considered for practical implementation like decision
period, overbooking and cancellation, arrival pattern of customers.
This concept can also be applied to other industry such as hotel industry, sea-cargo industry where
revenue needs to be high at very less expences.
8. 3. Aloysius George Research & Development Griantek, Chennai, India. Genetic
Algorithm Based Airlines Booking Terminal Open/ Close Decision System.
In this paper author given the information about revenue management of airline
company using ticket booking system with the help of genetic algorithm.
It is also stated that how prices of ticket changes time to time to maintain income level
constant or to higher level.
In given paper to operate flight booking terminal open-close system following data is
considered-
1. Historical Data Observation
2. Booking Frequency Generation Using Linguistic Variable
3. Booking Frequency Deviation
After considering this data Genetic algorithm is get applied in which firstly Population
Generation and Chromosome Representation is get done then fitness evaluation i.e
fitness value of each individual is get calculated then operation crossover is get applied
follows by mutation.
After mutation process gets terminated at particular point. As in this case decision get
fixed whether the terminal need to be open or closed at the particular time interval.
The main contributions of the paper is given as follows,
1. Incorporate the genetic algorithm for generating the optimal and suitable strategy
of airlines booking terminal openclose system
2. Analyze the past years observed booking data using Genetic Algorithm(GA).
9. CASE STUDY
Problem statement – To maximize the profit of airline company using Genetic Algorithm.
In this system, an application is taken to book seats for the customers in the airlines. So there are
number of customers requesting for the seats reservation from different terminals. There is a
booking terminal which accepts the request of one customer among various requesting customers;
requesting at the same time. The request from the terminal is accepted which contributes maximum
to the revenue, so as in the profit also. In this way, the profit can be maximized.
As shown in above block diagram method of yield management is get followed to get maximum
revenue.
The important data is categorized in various factors as follows-
1. Terminal- There can be any number of terminals which can raise the request for booking the
seats in the aircraft for the particular flight. Customer can make a request from any terminal
at any place either far or near. The maximum limit of requested seats by a terminal at one time
should be fixed.
2. Seats-Total number of seats in the aircraft is fixed. Initially the remaining seats are equal to
the total number of seats. Number of seats will be decremented each time a request is accepted.
3. Customers: There can be number of customer classes. Customers can be categorized on the
basis of their priorities and the fare they can pay etc. For example there can be leisure
customers and business customers. It can be assumed that the business customers can pay
more.
10. 4. Time: The total number of booking period must be specified i.e. the customer can raise the
request for the reservation during this time period before the flight takes off. The total time
period must be divided into time slots.
5. Fare: For each requested seat the customer has to pay the fare. The fare can be different for
different customers depending upon the type of customer and the time of request.
6. Revenue: Initially the total revenue is 0. Total revenue will be increased each time a booking
is made. Target is to maximize the revenue by selecting best request among number of requests
made at same time.
Use of Genetic Algorithm-
The basic genetic algorithm is as follows:
Start-Genetic random population of n chromosomes (suitable solutions for the problem)
Fitness- Evaluate the fitness f(x) of each chromosome x in the population.
New population- Create a new population by repeating following steps until the new
population is complete.
Selection- Select two parent chromosomes from a population per their fitness (the better
fitness, the bigger chance to get selected).
Crossover- With a crossover probability, perform crossover to the parents to form a new
offspring (children). If no crossover was performed, offspring is the exact copy of parents.
Mutation- with a mutation probability, mutate new offspring at each locus (Position in
chromosome)
Accepting- Place new offspring in the new population.
Replace- Use new generated population for a further sum of the algorithm.
Test- If the end condition is satisfied, stop, and return the best solution in current population.
Loop- Go to step2 for fitness evaluation.
For an example of this following assumptions are made-
1) Total number of Terminals = 3
2) Types of customers = 2 (Business Customer, Leisure Customer)
3) Total booking period = 3 months
11. 4) Time slots = 4 (slot 1(1st month), slot 2(2nd month), slot 3( 3rd month, slot 4(few hours
before flight takes off))
5) Initial total revenue = 0
6) Total seats = 70
7) Initial remaining seats = 70
8) Maximum booking allowed to the customer at one time = 3
9) Minimum Booking for a month = 18
10) Maximum booking for a month = 30
11) Fare for the business customer and leisure customer are shown in table.
Now the front task is to book the airline seats. But the main task that runs behind parallel with
the front task is to maximize the profit while booking the seats. So according to the example
the following requests are generated to book the seats: Month1: Terminal1: customer type-
2, number of seats- 1 Terminal2: customer type-2, number of seats- 2 Terminal3: customer
type-1, number of seats- 1 Among these three requests, one terminal is selected as this request
is passed to the proposed system and proposed system find out the optimized result. The
system gives the output as a terminal which contributes maximum to the revenue. In this
example the second terminal is selected. The revenue generated by the three terminals are
1500, 4000 and 2000.The second terminal is selected out of the three terminals because it
contributes maximum in the revenue. Similarly the requests are accepted for every month until
all the seats are booked or the time period for booking is finished. Every time the system
selects the best terminal out of the requested terminal. In this way at last the system generates
the maximum possible revenue.
12. The fitness function that is used in genetic algorithm to find out the best terminal is revenue =
no. Of seats * fare Fare depends upon the customer type, time of request and terminal. For this
example the total revenue generated for the 70 seats is 168200 where as the revenue generated
by the simple booking system is 136500. The result can be shown by graph also:
Discussion –
Every business including airline industry always wants their profit should be high. this target
is achieved by using Genetic algorithm. As problem stated above as a result it is better to show
unavailability rather than to close the terminal. When no. of customers’ requests for seat
reservation then the tickets gets booked on those terminals where ticket costs more which
further increases the revenue of the airline company.