Sei sulla pagina 1di 7

Abstract

Keywords

Introduction

Methodology

The proposed methodology gives a step-wise way to find out the desired combination.

1. Forecast of Load Factor

We have considered Load Factor a variable of time here. By applying regression analysis on the load factor
data of last 15 years, we found the following equation for calculating load factor for a particular year-

75.914 + 0.41375 (𝐶𝑢𝑟𝑟𝑒𝑛𝑡 𝑌𝑒𝑎𝑟 − 2001)

2. Calculation of Average Ticket Price

There are usually two classes of air seats- Business and Economy. However, considering the variety of
ticket prices offered, we have divided the seats into five categories- Premium, Business, Regular, Economic
and Promotional. Considering their proportion of the total seats, the average ticket price is found from
the following equation-

𝑎𝑅𝑝 + 𝑏𝑅𝑏 + 𝑐𝑅𝑟 + 𝑑𝑅𝑒 + 𝑒𝑅𝑝𝑟

Where,

a, b , c, d and e are proportion of the total seats occupied by Premium, Business, Regular, Economic and
Promotional seats respectively. And 𝑅𝑝 , 𝑅𝑏 , 𝑅𝑟 , 𝑅𝑒 𝑎𝑛𝑑𝑅𝑝𝑟 are the ticket prices of those seat categories.

3. Calculation of Competitive Pricing

To be on par with the existing competitor airlines, the data of ticket prices in different routes have been
collected. And with the help of regression analysis, ticket price was expressed as a function of flight
distance through two equations-one for short haul and another for the long haul:

𝑅 = 0.262𝐷 + 10.886 (Short Haul)

𝑅 = 0.088𝐷 + 143.646 (Long Haul)

4. Calculation of Revenue

With the estimation of average revenue and load factor complete, we now move to calculation of total
revenue from a flight. The following equation expresses that-

𝑅𝑖𝑗 = 𝑅 × 𝐿𝐹 × 𝑠𝑗

Here,
s is the total number of seats in the aircraft

5. Calculation of Flight Operating Cost

Cost of operating a flight has been estimated using the equation provided by Adler and Swan. The
equations express cost as a function of flight distance and number of seats in the aircraft-

𝑐𝑖𝑗𝑠ℎ𝑜𝑟𝑡 = (𝑑𝑖 + 722) × (𝑠𝑗 + 104) × 0.019 (Short Haul)

𝑐𝑖𝑗𝑙𝑜𝑛𝑔 = (𝑑𝑖 + 2200) × (𝑠𝑗 + 211) × 0.0115 (Long Haul)

6. Calculation of Revenue per Assignment

With cost and revenue known for each route and aircraft, we can now calculate profit in each route-
aircraft combination. The formulation for both short and long route should be-

𝑃𝑖𝑗𝑠ℎ𝑜𝑟𝑡 = (0.262𝐷 + 10.886) × 𝐿𝐹 × 𝑠𝑗 − (𝑑𝑖 + 722) × (𝑠𝑗 + 104) × 0.019

𝑃𝑖𝑗𝑙𝑜𝑛𝑔 = (0.088𝐷 + 143.646) × 𝐿𝐹 × 𝑠𝑗 − (𝑑𝑖 + 2200) × (𝑠𝑗 + 211) × 0.0115

7. Optimization with Ant Colony Algorithm

Now, we can focus on our prime target, which is to optimize the assignment of aircraft to different routes
in order to maximize profit. The formulation for the problem can be written as below-

For Short Haul:


𝑚 𝑛

𝑀𝑎𝑥𝑖𝑚𝑖𝑧𝑒 ∑ ∑[(0.262𝐷 + 10.886) × 𝐿𝐹 × 𝑠𝑗 − (𝑑𝑖 + 722) × (𝑠𝑗 + 104) × 0.019]𝑥𝑖𝑗


𝑖=1 𝑗=1
𝑚 𝑛

𝑠. 𝑡. ∑ ∑ 𝑥𝑖𝑗 = 𝑚
𝑖=1 𝑗=1
𝑛

∑ 𝑥𝑖𝑗 = 1
𝑗=1

𝑥𝑖𝑗 = 1 𝑜𝑟 0

𝑑𝑖 < 1000
For Long Haul:
𝑚 𝑛

𝑀𝑎𝑥𝑖𝑚𝑖𝑧𝑒 ∑ ∑[(0.088𝐷 + 143.646) × 𝐿𝐹 × 𝑠𝑗 − (𝑑𝑖 + 2200) × (𝑠𝑗 + 211) × 0.0115]𝑥𝑖𝑗


𝑖=1 𝑗=1
𝑚 𝑛

∑ ∑ 𝑥𝑖𝑗 = 𝑚
𝑖=1 𝑗=1
𝑛

∑ 𝑥𝑖𝑗 = 1
𝑗=1

𝑥𝑖𝑗 = 1 𝑜𝑟 0

𝑑𝑖 ≥ 1000
In both cases, i denotes the routes while j denotes the available aircrafts. Furthermore, m is the number
of flights whereas n is the total number of aircraft models. The decision of assignment xij is binary. And
the constraints ensure that we assign only one aircraft to one route.

8. Validation with Hungarian Algorithm

Numeric Illustration

For illustration, we have chosen 10 routes and 10 aircrafts, 5 for each of long and short haul trips. Their
distance and seat capacity are given below:

Let us first estimate the load factor. Applying 2017 in the formula we get the load factor to be-
75.914 + 0.41375 (2017 − 2001)
= 82.5
Next, the proportion of different aircraft seats are determined. We are using the following values-

a b c d e
0.04 0.06 0.4 0.3 0.2

Using these values, the average ticket price of the 10 routes were determined-
And from these data, competitive ticket pricing were calculated.

Now for applying ACO, we put the inputs for the two problems like this-

Problem 1 Problem 2
And the following parameters were set-
Finally, we also use the Hungarian algorithm to check whether our assignment is optimum. We can
calculate it through an online Hungarian Algorithm Calculator with providing the profit matrices. The
profit matrix for the two sets of problems are given below-
Result:

Optimal Assignment for Problem 1 is found below-

And Optimal Assignment for Problem 2 is found here-


And the result from the Hungarian algorithm is shown below:

As the results match, we can be sure that the assignment is optimum and profit has been maximized.

Conclusion:

PROFIT MAXIMIZATION of total supply chain profit AND minimization OF BULLWHIP EFFECT IN a multi-
echelon SUPPLY CHAIN network using particle swarm optimization and genetic algorithm.

International Journal of Integrated Supply Management Volume 11, No-2/3, 2017 Page 236-263

2016----Engin Pekel, Selin Soner Kara,

Simulation based fleet scheduling in the metro bus

International journal of simulation and process modelling volume -11, no-3/4, page-326 to 336, 2016

Ieee jounal of engineering management transaction

Potrebbero piacerti anche