Remote Tower Centre - Configuration and Planning of the Remote Tower Modules

LiU-ITN-TEK-G--013/077--SE

Remote Tower Centre

-Configuration and Planning of

the Remote Tower Modules

Peter Axelsson

Jonas Petersson

LiU-ITN-TEK-G--013/077--SE

Remote Tower Centre

-Configuration and Planning of

the Remote Tower Modules

Examensarbete utfört i Logistik

vid Tekniska högskolan vid

Linköpings universitet

Peter Axelsson

Jonas Petersson

Handledare Valentin Polishchuk

Examinator Tobias Andersson Granberg

Upphovsrätt

Detta dokument hålls tillgängligt på Internet – eller dess framtida ersättare –

under en längre tid från publiceringsdatum under förutsättning att inga

extra-ordinära omständigheter uppstår.

Tillgång till dokumentet innebär tillstånd för var och en att läsa, ladda ner,

skriva ut enstaka kopior för enskilt bruk och att använda det oförändrat för

ickekommersiell forskning och för undervisning. Överföring av upphovsrätten

vid en senare tidpunkt kan inte upphäva detta tillstånd. All annan användning av

dokumentet kräver upphovsmannens medgivande. För att garantera äktheten,

säkerheten och tillgängligheten finns det lösningar av teknisk och administrativ

art.

Upphovsmannens ideella rätt innefattar rätt att bli nämnd som upphovsman i

den omfattning som god sed kräver vid användning av dokumentet på ovan

beskrivna sätt samt skydd mot att dokumentet ändras eller presenteras i sådan

form eller i sådant sammanhang som är kränkande för upphovsmannens litterära

eller konstnärliga anseende eller egenart.

För ytterligare information om Linköping University Electronic Press se

förlagets hemsida

Copyright

The publishers will keep this document online on the Internet - or its possible

replacement - for a considerable time from the date of publication barring

exceptional circumstances.

The online availability of the document implies a permanent permission for

anyone to read, to download, to print out single copies for your own use and to

use it unchanged for any non-commercial research and educational purpose.

Subsequent transfers of copyright cannot revoke this permission. All other uses

of the document are conditional on the consent of the copyright owner. The

publisher has taken technical and administrative measures to assure authenticity,

security and accessibility.

Abstract

Today, many small aerodromes have a hard time surviving economically, and amongst the largest cost is air traffic control. Airlines are cutting costs where they can, and many times this affects the aerodromes as well, e.g. when airlines decide to park remotely instead of at the gate. The project called Remotely Operated Towers, initiated by SESAR and run by Saab and LFV, is aiming to address this problem.

The project revolves around remotely providing ATS to aerodromes where it is deemed suitable. A big challenge in this project is how to assign aerodromes to remote tower modules in the remote control centre. There are many ways to do this, but there is only a few ways to do it to achieve the least amount of modules.

This thesis aims to find an optimal solution to the challenge mentioned above.

The thesis resulted in a model where the user can provide the input of choice, i.e. aerodromes with associated ATS operating hours and movements, for a specific period – and receive the

assignment schedule for the modules, saying exactly which aerodrome are to be controlled by which module at what time.

Preface

This bachelor thesis is the completion of our education to become Air Traffic Controllers in Air traffic and Logistics at the Department of Science and Technology, Linköping University, during the fall/winter of 2013. The thesis comprises 16 ECTS and was written at Saab.

We would like to thank our supervisors at Saab, Bengt-Arne Skoog, and Mattias Johansson, for their help and commitment and for inviting us to interesting and rewarding simulations and study visits. We would also like to thank our supervisor Valentin Polishchuk and supervisor/examiner Tobias Andersson Granberg for their continuous support throughout the thesis work.

Special thanks to Håkan Brobeck at Transportstyrelsen for providing us with the statistics we asked for, and the air traffic controllers at Göteborg Landvetter Airport, who were kind enough to show us around the facilities and let us observe them while they participated in simulations.

Table of Contents

1 Introduction ... 1

1.1 Background ... 1

1.2 Short Description of Saab AB ... 2

1.3 Purpose ... 2

1.4 Limitations and Assumptions ... 2

1.5 Problem Description ... 2

1.6 Method ... 2

1.6.1 Gathering Information about the ROT Concept ... 3

1.6.2 Mathematical Optimization ... 3

1.6.3 Retrieving Data ... 4

1.6.4 Sources ... 4

1.7 Structure of Report ... 5

2 Theory ... 7

2.1 Current ATM System ... 7

2.2 Optimization ... 8

2.2.1 Simplex ... 8

2.2.2 CPLEX ... 8

2.3 Mathematical Modelling ... 8

2.3.1 Assignment and Knapsack Problem ... 9

2.3.2 Subset-Sum Problem ... 10

2.3.3 Bin-Packing Problem... 10

3 Remote Tower Concept ... 13

4 Assignment Model for RTMs ... 17

4.1 Input ... 17

4.2 Output ... 17

4.3 Constraints ... 17

5 Integer Programming Formulation ... 19

5.1 Mathematical Form ... 19

5.2 Verification ... 20

6 Data Collection ... 23

7 Results and Analysis ... 25

7.1 Different Results Depending on Period Size ... 25

7.4 ATS Operating Hours ... 30

8 Conclusion ... 31

9 Future Work ... 33

References ... 35

Appendices

Appendix B. AMPL input (2013-10-14) ... 39

Appendix C. Model in AMPL ... 43

Appendix D. Division of Active Aerodromes and Movements ... 45

Appendix E. Maximum number of movements per aerodrome ... 47

List of Figures

Figure 1. A schematic overview of a RTC ... 13

Figure 2. A possible configuration of two remote tower modules ... 14

Figure 3. Maximum amount of aerodromes per remote tower module ... 15

Figure 4. Example of rows in raw data... 23

Figure 5. Movements vs. active aerodromes ... 26

Figure 6. Model output, number of remote tower modules needed ... 28

Figure 7. Schedule for aerodromes to be handled by RTM9 ... 28

Figure 8. Number of movements at aerodromes in RTM9 ... 28

List of Tables

Table 2. Example run 1 - assignment ... 21

Table 3. Example run 2 - assignment ... 21

Table 4. Example run 3 - assignment ... 21

Table 5. All aerodromes that provides ATS in Sweden ... 24

Table 6. Number of aerodromes excluded depending on the size of the period ... 25

Table 7. Rough estimates for all weekdays ... 27

Table 8. Included and excluded aeodrome for October 14, 2013 ... 27

Table 9. Assignment of aerodromes to modules ... 29

Table 10. Number of RTMs needed for the week of 14th to 20th of October ... 29

Table 11. The difference in periods of providing ATS for 2013-10-14 ... 30

List of Equations

Equation 2.2. Knapsack problem ... 9

Equation 2.3. Subset-sum problem ... 10

Equation 2.4. Bin-packing problem... 11

Equation 5.1. Minimization of the RTMs used ... 19

Equation 5.2. ATS operating hours must be covered ... 19

Equation 5.3. All movements has to be handled ... 19

Equation 5.4. One aerodrome is only allowed to be controlled by one module at the time ... 19

Equation 5.5. One aerodrome is only allowed to be controlled by one module during all periods 19 Equation 5.6. The number of movements at a module must not exceed its maximum ... 20

Equation 5.7. The number of aerodromes at a module must not exceed its maximum ... 20

Equation 5.8. Movements must be handled, even outside ATS operating hours ... 20

Equation 5.9. Relationship between activePeriod and activeAD ... 20

Glossary

The following naming conventions are used in this thesis.

When referring to the control tower building the full word “tower” will be used, and for Remote

Tower Module, the term “module” will be used.

Aeronautical information publication (AIP) is a publication issued by or with the authority of a State and containing aeronautical information of a lasting character essential to air navigation. Approach control service (APP) is air traffic control service for arriving and departing flights. Air traffic control service (ATC) is the service provided to prevent collisions between aircraft, and on the maneuvering area between aircraft and obstructions; and expediting and maintaining an orderly flow of air traffic. It is divided into three subdivisions: APP, ACC (Area Control Service) and Aerodrome Control Service (TWR).

Air traffic control officer (ATCO). The person providing air traffic control service.

Air traffic management (ATM). The dynamic, integrated management of air traffic and airspace including air traffic services, airspace management, and air traffic flow management.

Air traffic services (ATS). Generic term for three services: Flight Information Service (FIS), Alerting Service (ALRS), and ATC.

A controller working position (CWP) is the ATCO workstation, including necessary ATS systems.

H24 - 24 hours a day.

Out-the-Window (OTW) view is as the name states, the view out the windows. The visual reproduction within the remote tower module is also described as OTW.

Remotely Operated Tower (ROT) is a concept of providing ATS for an aerodrome from a remote location. An aerodrome is considered as ROT-compatible if there is no more than a set amount of movements per period. This restriction is only an assumption.

Remote tower centre (RTC). A building which contains one or several Remote Tower Modules (RTMs). It can also include one or several RTM-detached controller working positions (CWPs). Remote Tower Module (RTM). A module including both the controller working position(s) (CWP) and the visual reproduction display.

1

Introduction

This report is the result of a bachelor thesis, written at the Institute of Technology (LiTH) at Linköping University (LiU), Norrköping, Sweden, during the fall/winter of 2013. The title of the

thesis is ‘Remote Tower Centre - Configuration and Planning of the Remote Tower Modules’.

1.1

Background

In the world we live in today, travelling by air is becoming as common as taking the train or the bus. Many people travel from less populated areas to the big cities for work and with the possibility to fly, the smaller airports are becoming more important for the cities/regions.

Even though the need of smaller airports is still there, it is hard to make them profitable. One large cost is for the air traffic control service. To be able to keep the small airports running, the industry needs to find a more cost-effective solution.

Today, every aerodrome with air traffic services (ATS) is controlled from a tower situated at the aerodrome. The air traffic controller(s) (ATCO(s)) in the tower is responsible for the traffic on the maneuvering area (runways and taxiways) and the traffic in the vicinity of the aerodrome (ICAO, 2007). The main mean of surveillance used by an ATCO is looking out the windows. In some towers, primary or secondary surveillance radar (PSR/SSR) and surface movement radar (SMR) is also used as means of surveillance. The remote tower concept aims to replace the local tower with cameras and sensors, while still providing the same service.

The Single European Sky ATM Research (SESAR) programme is a project for research and development launched by the European Union to meet the future need of capacity and safety within aviation (SESAR, 2013).

Saab AB is in charge of projects within SESAR that concerns the conduct of remote air traffic service (ATS) for one or more airports simultaneously – Remotely Operated Tower (ROT). By implementing the ROT concept, the cost of having air traffic controllers on duty and a building to maintain could be split between several airports (NORACON, 2013).

The concept of ROT is applicable to lots of aerodromes around the world, but today, neither Saab nor LFV has yet researched exactly how to implement it to be as efficient as possible. The focus from Saab and LFV has been to create a technical solution that works in practice (Skoog, 2013). The ROT concept is currently being evaluated in Sundsvall where a Remote Tower Centre (RTC) is located. From the centre, which is the only one in Sweden, Örnsköldsvik Airport will be controlled from one Remote Tower Module (RTM). In the RTC, a separate controller working position (CWP) is needed for the provision of approach control (APP) (Skoog, 2013).

Following the evaluations in Sundsvall, Saab will enter a phase of development where they need to test the system on a larger scale. The RTC in Sundsvall consists of only one module, but simulations including several modules and maybe even several RTCs must be performed in order to fully understand the effects and impact of a possible implementation of the concept.

Simulations with more than three aerodromes at the same time have not yet been performed, nor has any specific research as to how a simulation like that would be executed (Skoog, 2013).

1.2

Short Description of Saab AB

Saab AB is a Swedish aerospace and defense company founded in 1937. Saab has about 14,000 employees all over the world, with their focus on Europe and Sweden (Saab, 2013).

Saab offers world-leading solutions for military defense and security. The focus on research and

development is important for Saab, to stay ahead of the competitors and to meet the customers’

needs. This is accomplished by continuously adapting, developing, and improving the technology needed in different areas (Saab, 2013).

This thesis was performed in the business area called Security and Defense Solutions, one of six in total. The other five are Aeronautics, Dynamics, Electronic Defense Systems, Support and Services and the independent sister company called Combitech (Saab, 2013).

Everything from prison security to energy production is handled within the Security and Defense Solutions area. This thesis focuses on the air transportation and airport security field and the project called ROT (Saab, 2013).

1.3

Purpose

The purpose of this thesis is to find how to optimally assign aerodromes to remote tower modules in order to find the least amount of modules needed to provide remote air traffic services.

1.4

Limitations and Assumptions

For the problem to fit the scope of the thesis, some limitations and assumptions have to be made. This chapter will describe the limitations and assumptions, and why they are made.

For the purpose of this thesis, the concept of ROT and all its components are considered available and ready to be implemented wherever suitable. No emphasis will be placed on the technical aspect of the system, whether it functions correctly or not. The system is also considered reliable and no safety margin, i.e. for unusual events, is added. The assumptions are made to be able to create a model without taking any irregularities into account. A general limitation will be set as to decide which aerodromes that are or are not suitable to be remotely controlled.

The aspect of staffing the RTC, i.e. who works where when in the RTC, the need of separate positions for approach control, as well as where to locate the RTC will not be considered in this thesis.

1.5

Problem Description

The problem to assign aerodromes to remote tower modules needs to consider both operating hours as well as movements on the aerodromes, while finding the best combination according to the constraints given.

1.6.1 Gathering Information about the ROT Concept

To understand in what ways it would be technically possible to configure a remote tower module, information regarding the system was acquired. At Göteborg Landvetter Airport, Saab has

installed a camera tower and fitted a room with screens showing the picture captured by the cameras. In this room, some of the aerodrome’s most basic ATS equipment is installed (SMR, radio, and strip board). The purpose of this installment is to simulate a real environment in which the system has been implemented as a contingency measure.

The simulations are performed in “shadow-mode”, meaning that one or two ATCOs will be seated in the simulation room and listen to their colleagues in the actual control tower working the traffic, and try to work as they usually would using the means available to them, e.g. the usual technical aids, visual reproduction, and a special camera used as a binocular. These simulations were considered as very valuable in terms of retrieving background information about the concept, as well as engaging in discussions with the engineers at sight. From this experience, greater understanding of the concept and system was procured.

Most of the engineers working with ROT are situated in the headquarters in Växjö, and amongst them our supervisors. At the headquarters more advanced simulators has been installed. These are more elaborate and are able to show more than one aerodrome, as opposed to the one at Göteborg Landvetter Airport. As well as being able to show more aerodromes, the simulators also feature a digital strip board. The simulators in Växjö gave us the information needed regarding how the system, in the eyes of the engineers, will work and a basic idea of how a remote tower centre will work in practice.

1.6.2 Mathematical Optimization

This investigation will be based on a model into which the user should be able to input any aerodrome(s), with associated operating hours and traffic pattern, and be given the amount of remote tower modules needed in the remote tower centre.

The mathematical model was created after reading literature focusing on knapsack problems of different kinds. The problem of this thesis was finally considered to be a bin-packing problem. This is further explained in chapter 2.3.

At first, the focus was to define the different constraints and out of that create decision variables as well as parameters. This was made by looking at the problem stated and finding the different restrictions of the system. Decision variables was needed for all the restrictions, i.e. for maximum amount of movements/aerodromes for one remote tower module and for the periods specified. Some example data was created to make sure that the mathematical model was correct, according to the different parameters given, before writing it for a computer to solve.

The problem is solved by writing it in the modelling language called AMPL, which is later processed by the software with the same name and the solver called CPLEX. AMPL is often used for large-scale optimization problems, which the problem at hand is (AMPL Optimization LLC, 2004). Appendix A shows a general run file that can be used.

1.6.3 Retrieving Data

To be able to achieve fair and realistic results from the model, a lot of data needs to be acquired. The most accessible and most accurate data was to be found at EUROCONTROL. Upon request, they provided historical flight data and even forecasted flight data.

The data consists of flight plans regarding the enlarged ECAC1 area, which covers Europe. The data is split up into segments, each representing parts of that specific flight. Since the thesis focuses on departures and arrivals at aerodromes, all segments in between are unnecessary. To derive the necessary information, a decoding program was used to extract the time and aerodrome for departure and arrival respectively for each flight.

Information about Swedish aerodromes was retrieved from the Aeronautical Information

Publication (AIP) online. That includes the ATS operating hours, which is an essential part in the model. Statistics for the traffic (number of movements) at the specific aerodromes was received from Transportstyrelsen, the Swedish Transport Agency.

1.6.4 Sources

The primary sources for this thesis are literature and data. The literature have in most cases been found online, but retrieved in the original version where possible. Saab and its cooperators in SESAR are the founders of the ROT concept, and that makes the information hard to put in relation to anything else. Although, the source is considered reliable with reference to the thorough investigations made in this field.

Literature about optimization has been chosen carefully to fit our purpose. With some different opinions on the same subject, the sources were deemed credible.

Data has been retrieved from the national AIP (which LFV is responsible for), from

EUROCONTROL as well as Transportstyrelsen. All these sources are considered trustworthy, since they are government organizations and constantly being monitored by the authorities and media.

1.7

Structure of Report

The report begins with an introduction to the subject, background information, problem description, and a brief overview of the methods used throughout the thesis work.

In the second chapter, the theories behind the current ATS system as well as optimization are described in further detail, including the types of problems used to define the model.

The concept of ROT is described thoroughly in chapter 3, to explain the different parts of the system.

The model with input, output and constraints are presented in chapter 4, followed by the mathematical formulation derived from it in chapter 5, and a verification of the model is done. Following the need of actual data, chapter 6 describes the work of collecting it and how it was transformed to suit the model.

Chapter 7 presents the results together with some analysis. The report is finally summed up with a conclusion and some comments about future work to be done within the subject.

2

Theory

To address the problem as efficiently as possible some theoretic framework has to be established. A brief introduction to the current ATM system, i.e. the conventional tower, followed by

discussions regarding mathematical framework/models and optimization theory relevant to this thesis.

2.1

Current ATM System

The essentials of the current system are the physical tower located at the aerodrome and one or more ATCO(s) situated in the tower. From the tower the ATCO can see and monitor the maneuvering area to assure safe and orderly flow of traffic on and in the vicinity of the

aerodrome. The ATCO is also responsible for clearance delivery, ground control, managing traffic to and from the aerodrome, and flight data processing. In some cases, the TWR also provides Approach Control, in which case a separate working position with separate radar is fitted in the tower (ICAO, 1992).

The ATCO on sight is available to perform other tasks than the provision of ATS. The ATCO can perform tasks such as runway inspections or observe and report the local weather without any technical aids. However, these practices (runway inspections in particular) are performed by other personnel, such as someone responsible for the overall maintenance on the aerodrome. According to ICAO doc. 9426 Part III (1992) the aerodrome tower has to fulfil some requirements in order to properly control aircraft operating on and in the vicinity of the aerodrome:

A. the tower must permit the controller to visually survey those portions of the aerodrome and its vicinity over which he/she exercises control

B. the tower must be equipped so as to permit the controller rapid and reliable communications with aircraft with which he or she is concerned

Moreover, the requirements state that the controller must be able to distinguish between aircraft and vehicles operating on the same or different runways and/or taxiways. The height of the tower should be so to allow the controller to abide to previously mentioned requirements (ICAO, 1992). The ATCO use many means and systems to provide ATS. Some of them are optional, such as Surface Movement Radar (SMR), but the most important and most distinguishing mean for a local TWR, is the out-the-window (OTW) view (ICAO, 1992). Other examples of systems used for the provision of ATS:

 Air-ground communications, e.g. radio.

 Flight plan and ATS message handling systems  Light controls, e.g. for runway/taxiway lights  Binoculars

2.2

Optimization

Optimization (or mathematical optimization) revolves around finding a solution that satisfies the constraints given. The solution is optimal if it is feasible and the value is greater than every other feasible solution. This was an example of maximization but it works in a similar way for

minimization, with the goal to find the lowest feasible value. 2.2.1 Simplex

The simplex algorithm (or simplex method) is a method within optimization used to solve linear problems. According to the fundamental theorem of linear programming (LP) the optimal solution (maxima or minima) of a linear function over a convex polygonal region is always found at the

region’s corners. One of these solutions that is found in the corners of the region corresponds to

one basic feasible solution. The simplex method searches through all these basic feasible solutions in order to find an improvement of the objective function (Efstathiou, 2013).

Since the feasible region is convex, a basic feasible solution from which (neighbouring corners) no improvements to the objective function can be found is the optimal solution (Efstathiou, 2013). 2.2.2 CPLEX

CPLEX is an optimization software package that is named after the simplex method and

programmed in C. Robert E. Bixby is the founder of the program, but later on, it was acquired by ILOG and subsequently by IBM (IBM, 2013).

CPLEX solves integer programming problems and large LP-problems by applying the simplex method. There are several modeling systems from which this optimizer is accessible, e.g. AMPL (IBM, 2013).

2.3

Mathematical Modelling

Mathematical modelling is a way to describe the world as we see it, in mathematical terms. A model can help understand a problem, and how a solution might be procured. A model can take many forms, e.g. theoretical models or mathematical models (Sutharssan, 2011).

In this case, with aerodromes and different remote tower modules to handle them, the problem can be classified as an assignment or scheduling problem. The simplest version is called a knapsack problem, which can be applied to the problem of assigning the ATS operating hours at the aerodromes to the modules. However, when considering the amount of movements on each aerodrome, the problem extends to a bin-packing problem. In this chapter, some background information regarding the different approaches to the optimization part of the problem is stated.

2.3.1 Assignment and Knapsack Problem

The simplest and most intuitive description of the assignment problem is that of Hanan and Kurtzberg (1972). They describe it as optimally assigning men to jobs. Given the cost of assigning man to job , find an assignment of one man to one job in such a manner so to minimize the cost (Equation 2.1).

∑

Equation 2.1. Assignment problem

where , and is all the permutations of and is job assignment of person . Subsequently, the amount of ways to assign men to jobs is .

If, however, there is only one man available but many jobs to be done, one can see the connection to what is called the knapsack problem. Knapsack problem is defined as the problem of packing a knapsack with things, with the aim to fill up the knapsack as much as possible. The knapsack problem is thoroughly researched due to it often being part of more complex problems (Martello & Toth, 1990).

In the example stated above, one man represents one knapsack and one job represents one item. Variables and denotations for the general formulation of the knapsack problem are as follows. Va r ia bles { Denota tions

In Equation 2.2, a general formulation of the knapsack problem is shown.

∑

∑

{ }

Equation 2.2. Knapsack problem

The objective is to maximize the profit of the selected items. This should be done without the weight of the items exceeding the capacity of the knapsack (Lundgren, et al., 2003, p. 330).

2.3.2 Subset-Sum Problem

When the profit equals the weight of the items , i.e. the profit is the space used in the knapsack; the problem is to find a subset of items to fill up the knapsack but not to exceed its capacity. Equation 2.3 shows a general example of the subset-sum problem (Martello & Toth, 1990). ∑ ∑ { }

Equation 2.3. Subset-sum problem

The subset-sum problem mentions only one knapsack and maximizes the use of it, but since there

might be a need of several “knapsacks” (or remote tower modules); there is a need to extend it

into a multiple subset-sum problem, also known as the bin-packing problem (Martello & Toth, 1990).

2.3.3 Bin-Packing Problem

The bin-packing problem has one or several bins (“knapsacks”) with the same capacity. Instead of maximizing the use of one single bin, the bin-packing problem aims to minimize the number of bins used. The latter means that the bins chosen are used as efficiently as possible (Martello & Toth, 1990).

All items included in the bin-packing problem are assumed to have a positive weight, i.e. the bins cannot be enlarged, that does not exceed the capacity of one container (i.e. no item needs to be split between two or several containers). The assumption tells us that in the worst-case scenario,

number of items needs number of bins. This means that there is no bin with more than one

item inside, which also is the reason for the number of bins being equal to the number of items (Martello & Toth, 1990).

Va r ia bles

{

∑ ∑ { } ∑ { } { }

Equation 2.4. Bin-packing problem

The goal is to minimize the number of bins used, and at the same time not exceed the capacity of the bins used. One major thing that differentiate the bin-packing problem from the subset-sum problem and the knapsack problem is that all items needs to be placed in a bin (according to the second constraint in Equation 2.4) (Martello & Toth, 1990).

3

Remote Tower Concept

The objective of remote provision for single or multiple aerodromes is to provide the ATS defined in ICAO Annex 11, Documents 4444, 9426 for more than one aerodrome, by a single ATCO, from a remote location.

The concept is expected to be applied to low density2 to medium density3 aerodromes. The concept is not expected to be applied to larger aerodromes, e.g. Stockholm Arlanda Airport or Göteborg Landvetter Airport; however, the concept might be implemented as a contingency system (NORACON, 2013).

A RTC contains controller working positions (CWPs) and remote tower modules (RTM). The module is the “room” in which the visual reproduction display4 is located. Within the room, there can be one or more CWPs. For example another CWP can be fitted in the module if a certain aerodrome or event requires the attention of two ATCOs. However, the most common setup is one CWP in the module (NORACON, 2013).

Figure 1. A schematic overview of a RTC

Outside the module, but inside the RTC, separate CWPs can be installed. In the RTC showed in Figure 1 there are two additional CWPs, which in this case serves as a watch supervisor (WS) position that is responsible for the operations, and an approach control (APP) position. These positions are however optional and APP could instead be provided by the module. In larger RTCs, it is reasonable to assume that separate WS and APP positions would be installed (NORACON, 2013).

2 _{Mostly single movements, rarely simultaneous (e.g. at Kristianstad Airport)} 3

In the specific RTC in Figure 1, the approach position (CWP APP) is providing approach control to all four aerodromes, and the modules are providing TWR to the same aerodromes.

Figure 2. A possible configuration of two remote tower modules

Figure 2 shows the setup that will be further investigated. Several modules controlling several aerodromes. The investigation and analysis will be based on the module and what aerodromes are assigned to it at what time. No emphasis will be given to the amount of CWPs inside the module. To allow the transition between aerodromes to be as fast and efficient as possible the module will be required to have a unified layout. Today the layout, or Human Machine Interface (HMI), varies between different towers. The unified HMI in a module will eliminate the need to either change the HMI of the working position or require the controller to be licensed for a specific HMI at a specific aerodrome (NORACON, 2013).

Figure 3. Maximum amount of aerodromes per remote tower module

The out-the-window (OTW) view provided to the controller in a module is a reproduction of the actual view, using the means of cameras and sensors. Depending on how far away from the aerodrome the module is located (or rather the length of the Ethernet cable linking the two), there is some delay, i.e. what is shown is not actually real-time. However, this delay is usually not more than one second (Johansson, 2013).

The reproduced OTW view is solely based on a number of cameras mounted on a camera-tower at every aerodrome respectively, and if one or more of these cameras shows a false picture or no picture at all, the OTW surveillance would seize to exist. If no other surveillance equipment is available at the aerodrome (e.g. SMR), the service would be heavily reduced or even terminated. In a conventional tower, it is unlikely that the OTW view would be obstructed in a similar way.

4

Assignment Model for RTMs

The mathematical model that describes the problem is in this section divided into smaller parts, which are further explained. The first part is input where all the information needed besides the model itself is stated. Following the input chapter is output, here the desired output is described. Lastly, all the constraints and their properties are presented and explained.

4.1

Input

The input to the model contains five parts, each listed below.

 Set of aerodromes to be analyzed  Set of periods to be analyzed  Set of RTMs to be analyzed

 Maximum number of movements per RTM per period  Maximum number of aerodromes per RTM

 Number of movements at every aerodrome j during every period k

 Matrix representing ATS operating hours (0/1 for aerodrome j and period k) Input given to AMPL from the date of 2013-10-14 is found in Appendix B.

4.2

Output

The output is an assignment of aerodromes to RTMs.

4.3

Constraints

ATS oper a ting hour s must be cover ed

If an aerodrome provides ATS during a period, that aerodrome needs to be assigned to a RTM. All movements ha s to be ha ndled

Total number of movements at all RTMs for an aerodrome for each period must be equal to the number of movements at the aerodrome for the same period.

One a er odr ome is only a llowed to be a ssigned to one RTM a t the time Due to the extra work needed to manage the same traffic from two different RTMs. One a er odr ome is only a llowed to be a ssigned to one RTM dur ing a ll per iods To avoid unnecessary switching between RTMs, and to simplify staffing later on (Future Work). The number of movements a t a RTM must not exceed its ma ximum

The total amount of movements at one RTM must not exceed the maximum capacity. The number of a er odr omes a t a RTM must not exceed its ma ximum

The number of aerodromes assigned to one single RTM must not exceed the maximum capacity. Movements must be ha ndled, even outside ATS oper a ting hour s

5

Integer Programming Formulation

For the model described in the previous chapter to be implemented in AMPL, it needs to be formulated mathematically. Va r ia bles { { { In the variables above, as well as in the following descriptions of the constraints,

; ; and .

5.1

Mathematical Form

Equation 5.1. Minimization of the RTMs used

Subject to

∑

Equation 5.2. ATS operating hours must be covered

∑

Equation 5.3. All movements has to be handled

∑

Equation 5.4. One aerodrome is only allowed to be controlled by one module at the time

∑

∑

Equation 5.6. The number of movements at a module must not exceed its maximum

∑

Equation 5.7. The number of aerodromes at a module must not exceed its maximum

Equation 5.8. Movements must be handled, even outside ATS operating hours

The constraints displayed above each describes a constraint mentioned in chapter 4.3. For the sake of the model, the decision variables needs to be connected with constraints, as to be dependent on each other. The following constraints is created to keep the relationship between the different decision variables.

∑

Equation 5.9. Relationship between activePeriod and activeAD

∑

Equation 5.10. Relationship between activeAD and activeRTM

The constraints, together with the rest of the model, can be found in the AMPL model file, attached as Appendix C.

5.2

Verification

To verify that the model works as expected, some example data was created and provided as input. This data was created to check if the solutions from the model is feasible and to see if the constraints work as they are supposed to. All the examples are solved with AMPL/CPLEX. In this verification, is set to 6 and is set to 3. In Table 1, one column represent one period, and in this case one period is one hour. For easier readability, only three periods is included.

For the first run, with only period 1, one could assume that AD1 is assigned to RTM1. Since AD1 has 6 movements, neither AD2 or AD3 could be assigned to that RTM without exceeding the maximum number of movements allowed per period. Those two are instead assigned to a second module. Finally, both AD4 and AD5 does not have any movements, which means that they could be assigned to RTM1 together with AD1, without violating the constraint regarding maximum number of movements. The assignment could be seen in Table 2.

Table 2. Example run 1 - assignment

Modules Assigned aerodromes Period 1

RTM1 AD1, AD4, AD5 6

RTM2 AD2, AD3 5

When the second period is included, the model needs to consider both period one and two in the result. The first solution with AD1, AD4 and AD5 together is not possible, since that would result in eight movements in period two. If we only consider the second period, there is only one way to assign the aerodromes to two modules. As seen in Table 3, AD2, AD3 and AD4 are combined which is also feasible for the first period.

Table 3. Example run 2 - assignment

Modules Assigned aerodromes Period 1 Period 2

RTM1 AD1, AD5 6 6

RTM2 AD2, AD3, AD4 5 6

Finally, with all three periods included, the only solution for the first two periods is no longer feasible (AD1 and AD5 has seven movements in period three). That means that the model gives a solution with three modules, quite similar to the first solution, except for AD5 that is assigned to its own module (Table 4).

Table 4. Example run 3 - assignment

Module Assigned aerodromes Period 1 Period 2 Period 3

RTM1 AD1, AD4 6 5 3

RTM2 AD2, AD3 5 4 5

RTM3 AD5 0 3 4

According to this verification, the model works as we expected it to do and the solutions given was optimal (including feasibility).

6

Data Collection

In Sweden, as well as every other country in the world, there are one or many aerodromes in use every day. The major input to the model is movements at each aerodrome and that information was received from the Demand Data Repository (DDR) hosted by EUROCONTROL.

DDR2 is the second version of the Demand Data Repository, which has the objective to “provide European airspace planners and airspace users with an accurate picture of European air traffic

demand, past and future, that will meet their planning and monitoring needs”. The data in the

database is collected from the last filed flight plan with status ‘accepted’ that interferes with the enlarged ECAC area (EUROCONTROL, 2013).

The raw data file is a space-delimited text file and every flight segment is written on a separate row. Each flight is composed of a number of segments. For each turn, climb or descent, or the passing over a significant point - the segment is completed and a new segment is initiated. The number of segments for a whole flight varies. For the purpose of this thesis, the information regarding departure and arrival aerodrome along with departure and arrival times for each flight is desired. Rows in the data (from 2013-11-12) have the format seen in Figure 4 (irrelevant data phased out):

[...] CYYZ EGKK [...] 025845 051502 [...] 131112 131112 [...] 2831210289 [...] [...] CYYZ EGKK [...] 051502 051552 [...] 131112 131112 [...] 2831210289 [...]

Figure 4. Example of rows in raw data

The information shown in this specific snippet of data, from left to right, corresponds to:

 departure aerodrome (ICAO code)  arrival aerodrome (ICAO code)

 start time for this segment (HHMMSS)  end time for this segment (HHMMSS)  start date of this segment (YYMMDD)  end date of this segment (YYMMDD)

 flight ID (unique for each flight, but the same for each segment for that flight)

The data is considered reliable, it is however difficult to verify the data. Each data file from DDR2 represents one day of flights concerning European airspace, and one file is approximately 170 MB. To download and summarize all the data for a whole year and compare it with statistics published by Transportstyrelsen would be extremely time-consuming.

The data from EUROCONTROL includes traffic regarding all aerodromes in Europe, which means even heliports and small aerodromes without ATS. Transportstyrelsen has provided us with data concerning the aerodromes with ATS for each day in a whole week. This data is then used for verification. Comparisons between the data received from EUROCONTROL and the data provided by Transportstyrelsen show a slight difference, however for the purpose of this thesis the data from DDR2 is considered sufficient and reliable.

The national AIP (LFV A, 2013) and NOTAM5 (LFV B, 2013) for Sweden has been consulted to obtain the ATS operating hours for the aerodromes in Sweden (Table 5). Some aerodromes provides ATS upon request and other might be controlled by AFIS officers during parts of the day. Only the actual ATS operating hours has been retrieved. The ATS operating hours as well as the data from EUROCONTROL are all in UTC6 time.

Table 5. All aerodromes that provides ATS in Sweden

ESNX ARVIDSJAUR ESNS SKELLEFTEÅ

ESSD BORLÄNGE ESSA STOCKHOLM/Arlanda

ESGG GÖTEBORG/Landvetter ESSB STOCKHOLM/Bromma

ESGP GÖTEBORG/Säve ESKN STOCKHOLM/Skavsta

ESMT HALMSTAD ESOW STOCKHOLM/Västerås

ESGJ JÖNKÖPING ESNN SUNDSVALL-TIMRÅ

ESMQ KALMAR ESIB SÅTENÄS

ESIA KARLSBORG ESGT TROLLHÄTTAN/Vänersborg

ESOK KARLSTAD ESNU UMEÅ

ESNQ KIRUNA ESCM UPPSALA

ESMK KRISTIANSTAD ESPE VIDSEL

ESCF LINKÖPING/Malmen ESSV VISBY

ESSL LINKÖPING/SAAB ESMX VÄXJÖ/Kronoberg

ESTL LJUNGBYHED ESNZ ÅRE ÖSTERSUND

ESPA LULEÅ ESTA ÄNGELHOLM

ESMS MALMÖ ESOE ÖREBRO

ESSP NORRKÖPING/Kungsängen ESNO ÖRNSKÖLDSVIK

ESDF RONNEBY

In the AMPL-modelling, two instances of parameter matrices are needed: one for the total amount of movements and one for the ATS operating hours of the aerodromes. To create the matrices the data had to be processed in a way so that AMPL would be able to interpret it correctly.

With the use of a Python program, the data was processed and a text file was produced that was formatted in the desired way.

7

Results and Analysis

The purpose of this thesis was to find how to optimally assign aerodromes to remote tower modules in order to find the least amount of modules needed to provide remote ATS. The analysis and results are based on a week in 2013 (14th to 20th of October), and when focusing on a

specific day, October 14, 2013 is used.

Information about movements does only include arrivals and departures, and other movements such as touch-and-go landings and aircraft crossing the control zone is not taken into account. Even though forecasted flight data would be interesting to analyze, it is only available for two days ahead which does not give enough data for a thorough analysis. This is discussed further in Conclusion and Future Work.

7.1

Different Results Depending on Period Size

The aerodromes that are provided as input are those that do not exceed a set amount of

movements per period, i.e. are ROT-compatible. Depending on the size of the period (in terms of time), more or less aerodromes might be included/excluded. This is due to the fact that it is more likely that aerodromes have for example three movements in half an hour, than six movements in a full hour. This is demonstrated in Table 6, with statistics from the week used, where the

percentage is the amount of aerodromes excluded in relation to the total amount.

Table 6. Number of aerodromes excluded depending on the size of the period

Date Total AD Hours Half hours

2013-10-14 35 10 29% 18 51% 2013-10-15 35 11 31% 18 51% 2013-10-16 35 12 34% 20 57% 2013-10-17 35 12 34% 17 49% 2013-10-18 35 11 31% 18 51% 2013-10-19 35 11 31% 13 37% 2013-10-20 35 11 31% 17 49%

During the weekdays and Sunday (2013-10-20), splitting up by hours excludes on average one fourth of the compatible aerodromes and splitting by half hours results on average in a third of the aerodromes excluded. The results from the analysis for different number of periods is not suitable for comparison, since the aerodromes included in the analysis might differ.

Since the number of aerodromes deemed as ROT compatible differs so much depending on period size, it is probably not the best way to determine compatibility. Instead, a set amount of

aerodromes that are chosen independently of is preferable. However, by using the current method, the aerodromes that are interesting might be identified.

Splitting the periods in half will result in a lower number of aerodromes being ROT-compatible with the same relative number of movements. If the goal is to keep approximately the same number of aerodromes, the number of movements per half hour ( ) must be set to four instead of three.

7.2

Estimation and Analysis of Total Values

From now on, after what has been found in chapter 7.1, 24 periods are used (hours) and the maximum number of movements per module per period is six. As stated before, the number of aerodromes per module is restricted to three.

Figure 5 displays the number of movements and the number of active aerodromes for every period during the 14th of October. There are four obvious peaks in the traffic, whilst the number of aerodromes is somewhat stable. Overall, the average number of movements is greater than the number of active aerodromes (dashed lines), which in practice means that there is more than one movement per aerodrome per period.

Figure 5. Movements vs. active aerodromes

What has been noted is that there is two fast ways to find a rough estimate of how many modules are needed for the traffic for any period. The first is based on active aerodromes, and it is found by dividing the maximum amount of active aerodromes in that period with the maximum amount of aerodromes a single module can handle simultaneously. During the day shown in Figure 5, the maximum number of active aerodromes is 29 (periods 8, 9, 10, 13 and 14), which gives a

minimum number of 10 modules needed, since the value needs to be rounded up.

0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Total active aerodromes Total movements

As mentioned, these are two rough estimates, only considering number of active

aerodromes/movements and the limitation of a module. The number of modules used can never be less than the highest of the two estimates, but since the model has to consider both, the total number of modules needed could be greater. This example is only calculated out of one period, and if all periods are included, there might be limitations reached in other periods.

Appendix D shows the final division of movements and active aerodromes.

As stated in the previous chapter, it is important to consider the number of periods chosen when deciding on the constraint regarding movements per period. It is for example easier to find half hours with more than three movements than it is to find full hours with more than six. This is important since that in turn decides which aerodromes to include in the input.

In Table 7 the maximum number of aerodromes and movements per day is shown and the estimate derived from that number. In the right most column the actual number of RTMs according the model is presented.

Table 7. Rough estimates for all weekdays

Date Max aerodromes Rough estimate (AD) Max movements Rough estimate (mov) RTMs needed 2013-10-14 29 9,67 ≈ 10 53 8,83 ≈ 9 10 2013-10-15 27 9 58 9,67 ≈ 10 10 2013-10-16 25 8,33 ≈ 9 55 9,17 ≈ 10 10 2013-10-17 27 9 57 9,5 ≈ 10 10 2013-10-18 27 9 48 8 9 2013-10-19 17 5,67 ≈ 6 28 4,67 ≈ 5 9 2013-10-20 23 7,67 ≈ 8 43 7,17 ≈ 8 9

7.3

Results from the Model

In this chapter, the number of periods selected is 24 (one for each hour of the day) and the maximum number of aerodromes per period is three.

35 aerodromes from Sweden was introduced to the model, but with the constraint of the maximum number of movements per module per period set to six, only 29 was considered

“ROT-compatible”. A table showing which aerodromes that are considered ROT-compatible and those

that are not can be seen below (Table 8). A more extensive table also showing the maximum amount of movements at each aerodrome for October 14, 2013 can be found in Appendix E.

Table 8. Included and excluded aeodrome for October 14, 2013

Included Excluded

ESCF, ESCM, ESDF, ESGJ, ESGP, ESGT, ESIA, ESIB, ESKN, ESMK, ESMQ, ESMT, ESMX, ESNN, ESNO, ESNQ, ESNS, ESNX, ESNZ, ESOE, ESOK, ESOW, ESPA, ESPE, ESSD, ESSL, ESSP, ESTA, ESTL

ESGG, ESMS, ESNU, ESSA, ESSB, ESSV

When running the model with the input from the date specified above, the model returns the minimum amount of modules needed (Figure 6).

1 NumberActive 10

Figure 6. Model output, number of remote tower modules needed

An assignment schedule, displayed as a matrix, shows the intended usage of each module i.e. which aerodrome that has to be assigned to which module when. The matrix for RTM9 is found in Figure 7. activePeriod[9,j,k] : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 := ESGT 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 ESOW 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 ESTA 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0

Figure 7. Schedule for aerodromes to be handled by RTM9

Another matrix shows the number of movements handled at every period for the active aerodromes at one module. Figure 8 shows the matrix for RTM9.

movements[9,j,k]

: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 := ESGT 0 0 0 0 1 0 2 0 1 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 ESOW 0 0 0 0 2 1 1 0 2 4 2 0 5 4 3 3 1 3 5 2 1 0 0 0 ESTA 0 0 0 0 2 2 1 6 1 0 1 2 1 1 2 1 1 3 1 0 1 0 0 0

Figure 8. Number of movements at aerodromes in RTM9

The model makes the assignment of aerodromes with regard to the constraints mentioned before, e.g. six movements per hour and not more than three aerodromes per module. As seen in period 7, 12, 14, 17 and 18, this combination equals exactly six movements, distributed in different ways among the aerodromes.

The solution shows that for each one of the periods mentioned before, six movements needs to be handled. Even though there is 60 minutes in that period (i.e. 10 minutes per movement), there is a

During the day of October 14th, the assignment of aerodromes to modules would be as in Table 9.

Table 9. Assignment of aerodromes to modules

Module Assigned aerodromes

RTM1 ESMT, ESNO, ESSD

RTM2 ESDF, ESMQ, ESSL

RTM3 ESCF, ESKN, ESMK

RTM4 ESGJ, ESOK, ESSP

RTM5 ESIB, ESNQ, ESNZ

RTM6 ESNS, ESNX, ESPA

RTM7 ESGP, ESPE, ESTL

RTM8 ESCM, ESNN

RTM9 ESGT, ESOW, ESTA

RTM10 ESIA, ESMX, ESOE

To see the number of movements and operating hours (i.e. the assignment), the entire output can be found in Appendix F.

The result for all days in the week is shown in Table 10.

Table 10. Number of RTMs needed for the week of 14th to 20th of October

Date RTMs 2013-10-14 10 2013-10-15 10 2013-10-16 10 2013-10-17 10 2013-10-18 9 2013-10-19 9 2013-10-20 9

The amount of RTMs needed is almost the same during the week, but since there is a need for ten RTMs on Monday (2013-10-14) through Thursday (2013-10-17), a solution for the entire week must include at least that amount of RTMs.

7.4

ATS Operating Hours

If the ROT concept are to be implemented with the goal of providing ATS to all aerodromes 24

hours a day, then the total amount of “working periods” will increase compared to keeping the

current ATS operating hours. As seen in Table 11, the amount of periods where ATS is provided H24 increases by approximately 50% compared to the current hours (for 2013-10-14). This could

be interpreted as an increased “cost” of 50%. What is interesting, but not surprising, is that the

same amount of modules is needed. This in turn means that in the scenario where the current operating hours are kept, the modules have a lot of unused periods.

The calculation is made by multiplying the number of included aerodrome with 24 periods, in this case, .

Table 11. The difference in periods of providing ATS for 2013-10-14

Current ATS hours ATS H24

Periods providing ATS 472 696

Number of modules 10 10

The unused periods would only be profitable if it was possible to find a user who needs them. This could either be an airline with a new flight, or another aerodrome (preferably in another time

zone to fit the “free” hours). Although, the last alternative needs some changes to be made in the

number of aerodromes one module is allowed to control in total.

There is a possibility to apportion the current daily movements in a better way, as mentioned in chapter 7.3, to utilize the modules to their maximum.

8

Conclusion

To achieve the purpose of this thesis, an optimization model was created. The model creates an assignment schedule for each remote tower module needed for the aerodromes provided as input. The schedule shows which aerodromes should be assigned to what module at what time.

As the results are based on an optimization model, either a feasible solution or no solution at all is retrieved, due to the constraints. The complexity of the solution depends on how many

aerodromes that is included as well as the number of movements at each one of them.

What has been found is that the rough estimate mentioned in chapter seven is accurate most of the times. This method will, however, only result in the amount of remote tower modules needed, and not how to use them optimally (how to assign aerodromes to modules). This estimate could be used in a way to initially find how many modules that are needed in a RTC, but for the optimal assignment of aerodromes to the modules an optimization model is needed. The best use of the model in this thesis is to input data for a whole year to see how many modules the RTC would need, and for the operational use of the modules input forecasted data provided by

EUROCONTROL to find how many modules to “open” each day and how to use them.

Furthermore it has also been noted that the choice of the size of the periods is of great importance. Choosing smaller periods, with the same relative traffic limit as a larger period, results in more

aerodromes being excluded, according to the definition of “ROT-compatible” in this thesis.

In the analysis, the maximum amount of movements a single module can handle per period is set to six. In theory, this means that six movements could be split amongst three different

aerodromes. In reality, one ATCO’s total workload when controlling three aerodromes

simultaneously would probably be higher than for one aerodrome and the same amount of traffic. With the previous statement in mind, the capacity for one module would depend on how many aerodromes that are controlled simultaneously. However, this is not considered in the model. Before making the decision about abandoning the actual tower and control the aerodrome from a remote location, both operating hours and traffic must be considered thoroughly. The model presented in this thesis could, with some extensions, be a good basis for such a decision.

9

Future Work

This thesis only calculates the number of working positions needed to handle the aerodromes specified for the model during the period of traffic supplied. The next step would be to analyze the need during a longer period and out of that find a feasible setup of one or several RTCs. An analysis of for example a whole year would provide a more reliable result since a weekly or daily analysis might result in too few modules compared to the need for another week/day (see Table 10 on page 29). The traffic from year to year does not change much, except for the general increase of traffic for the whole market.

The current model maximizes the amount of movements for each module. What should be done is to optimize the workload across all modules. This means that one module should not have much more or less traffic than any other module, thus creating a more balanced workload for all modules. What can be further analyzed regarding the module is if it is possible to increase the capacity of any module by having an additional ATCO.

While the number of RTMs is decided, a lot of work is still to be done regarding staffing. Working hours, local agreements, and so on needs to be fulfilled, as well as the part of air traffic control called ratings. The biggest question is how to combine the aerodromes to make the ATCOs as efficient as possible with their ratings.

There might be a need to cluster the aerodromes in a specific way to make it easier for the ATCOs to handle the traffic. This can be done in several ways and since it has not been analyzed properly, no one knows which combination would be the best. Some examples of possible clustering forms are by geographical position, runway configurations, or amount of traffic.

The location of the RTC is also an important aspect that has to be researched. There could be advantages in placing a RTC in close proximity to an aerodrome which is not subject to remote control, and in that way be able to share the personnel. It could be beneficial to co-locate the RTC together with an air traffic control centre, where technicians, infrastructure, and facilities are already available.

References

AMPL Optimization LLC, 2004. AMPL FAQ. [Online]

Available at: http://www.ampl.com/FAQ/index.html#WhatisAMPL [Accessed 11 November 2013].

Efstathiou, C. E., 2013. Simplex Optimization. [Online]

Available at: http://www.chem.uoa.gr/applets/AppletSimplex/Appl_Simplex2.html [Accessed 22 December 2013].

EUROCONTROL, 2013. DDR2. [Online] Available at: http://www.eurocontrol.int/ddr2 [Accessed 16 November 2013].

Hanan, M. & Kurtzberg, J. M., 1972. A Review of the Placement and Quadratic Assignment Problems. SIAM Review, 14(2), pp. 324-342.

IBM, 2013. IBM CPLEX Optimizer. [Online]

Available at: http://www-01.ibm.com/software/commerce/optimization/cplex-optimizer/ [Accessed 22 December 2013].

ICAO, 1992. Doc 9426: Air Traffic Services Planning Manual. 1 ed. Chicago: International Civil Aviation Organization.

ICAO, 2001. Annex 11: Air Traffic Services. 13 ed. Chicago: International Civil Aviation Organization.

ICAO, 2007. Doc 4444: Air Traffic Management. 15 ed. Chicago: International Civil Aviation Organization.

Johansson, M., 2013. Technical aspects of the ROT concept [Interview] (2 October 2013). LFV A, 2013. IAIP, Norrköping: LFV.

LFV B, 2013. NOTAM Summary, Norrköping: LFV.

Lundgren, J., Rönnqvist, M. & Värbrand, P., 2003. Optimeringslära. 3:3 ed. Lund: Studentlitteratur.

Martello, S. & Toth, P., 1990. Knapsack Problems : Algorithms and Computer Implementations. Chichester: John Wiley & Sons.

NORACON, 2013. OSED for Remote Provision of ATS to Aerodromes, including Functional Specification, Brussels: SESAR Joint Undertaking.

Saab, 2013. Saab in brief. [Online]

Available at: http://saabgroup.com/en/About-Saab/Company-profile/Saab-in-brief/ [Accessed 26 September 2013].

SESAR, 2013. Background on Single European Sky. [Online] Available at: http://www.sesarju.eu/about/background

Skoog, B.-A., 2013. Current Approach [Interview] (3 December 2013).

Sutharssan, T., 2011. An Introduction to Mathematical Modelling. 1 ed. Greenwich: University of Greenwich.

Appendix A.

AMPL run file

# OPTIONS

option solver cplexamp; option presolve 1; option display_transpose -200; option display_width 150; option show_stats 1; option omit_zero_rows 1; option omit_zero_cols 0; option times 0;

# CONNECTIONS and SOLVE model mod.txt; data dat.txt; solve; # DISPLAY display mAD; display mMR;

display _objname, _obj;

display _solve_elapsed_time; # EXIT

Appendix B.

AMPL input (2013-10-14)

set AD := ESCF ESCM ESDF ESGJ ESGP ESGT ESIA ESIB ESKN ESMK ESMQ ESMT ESMX ESNN ESNO ESNQ ESNS

ESNX ESNZ ESOE ESOK ESOW ESPA ESPE ESSD ESSL ESSP ESTA ESTL ;

# ESGG ESMS ESNU ESSA ESSB ESSV ESGK ESGL ESGR ESHQ ESKM ESMO ESND ESNG ESNJ ESNK ESNL ESNV

ESOH ESST ESSU ESUP ESUT

param nR := 29 ; param p := 24 ; param mAD := 3 ; param mMR := 6 ; param num_mov: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 := ESCF 0 0 0 0 0 0 0 1 1 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 ESCM 0 0 0 0 0 0 0 0 0 1 1 0 1 2 0 0 0 0 0 0 0 0 0 0 ESDF 0 0 0 0 2 1 1 3 0 0 2 1 0 1 1 1 0 3 0 0 1 0 0 0 ESGG 3 0 2 1 13 17 9 12 12 9 11 8 8 5 14 18 13 15 8 10 11 4 1 3 ESGJ 0 0 0 0 1 1 0 2 1 1 0 0 1 1 3 1 2 1 0 0 1 1 2 1 ESGK 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 ESGL 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ESGP 0 0 0 1 0 1 2 6 5 2 2 4 3 3 4 3 6 4 2 3 0 0 0 0 ESGR 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 ESGT 0 0 0 0 1 0 2 0 1 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 ESHQ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 ESIA 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ESIB 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 ESKM 0 0 0 0 2 0 0 1 2 1 0 0 0 0 2 0 0 2 0 0 0 0 0 0 ESKN 0 0 0 0 3 3 5 3 2 4 1 3 3 3 4 1 3 5 3 4 2 3 1 1 ESMK 0 0 0 0 1 0 1 1 0 0 0 0 0 1 2 1 0 1 0 0 0 0 0 0 ESMO 0 0 0 0 0 0 3 1 2 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 ESMQ 0 0 0 0 2 0 1 3 0 1 0 2 2 2 1 2 2 2 0 0 1 0 0 0 ESMS 1 1 3 1 3 9 8 7 3 6 3 3 4 4 6 9 9 5 3 6 5 6 0 3 ESMT 0 0 0 0 1 2 0 3 1 0 0 1 0 1 4 0 1 1 0 0 0 0 0 0 ESMX 0 0 0 0 1 1 0 3 1 1 4 2 1 0 1 2 2 2 0 1 1 0 0 2 ESND 0 0 0 0 1 0 0 0 1 0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 ESNG 0 0 0 1 3 1 1 1 0 1 0 1 0 2 2 0 2 2 0 1 0 0 0 0 ESNJ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 ESNK 0 0 0 0 0 2 0 0 2 0 0 0 0 0 0 2 0 0 2 0 0 0 0 0 ESNL 0 0 0 0 2 0 0 0 1 1 0 0 0 0 1 1 0 0 2 0 0 0 0 0

ESNO 0 0 0 0 2 0 0 1 1 0 0 0 1 1 0 0 1 2 0 1 0 0 0 0 ESNQ 0 0 0 0 4 0 0 1 0 0 2 2 0 0 1 1 0 1 0 0 1 0 0 0 ESNS 0 0 0 0 1 0 0 2 0 2 1 1 2 0 0 0 1 1 0 0 1 0 0 0 ESNU 1 0 1 1 2 3 2 6 3 1 5 2 1 1 5 3 3 7 1 5 4 1 2 2 ESNV 0 0 0 0 1 1 0 0 1 1 0 0 0 0 2 0 0 0 1 0 0 0 0 0 ESNX 0 0 0 0 1 0 0 0 0 1 0 0 0 0 1 0 0 0 1 0 0 0 0 0 ESNZ 0 0 0 0 2 2 0 5 1 1 2 0 0 1 3 1 1 2 3 1 1 0 0 0 ESOE 0 0 1 0 1 1 0 1 2 0 0 0 0 0 2 2 0 4 2 0 0 1 0 1 ESOH 0 0 0 0 2 0 0 2 0 0 0 0 0 0 1 1 0 2 0 0 0 0 0 0 ESOK 0 0 0 1 1 0 0 1 2 0 0 0 0 1 2 0 1 2 0 0 1 0 0 0 ESOW 0 0 0 0 2 1 1 0 2 4 2 0 5 4 3 3 1 3 5 2 1 0 0 0 ESPA 0 0 0 1 4 4 3 3 4 2 1 3 2 2 3 6 4 5 2 0 2 1 0 0 ESPE 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ESSA 0 1 1 3 24 61 59 45 44 31 34 41 32 51 48 60 57 32 42 40 23 24 6 6 ESSB 0 0 0 0 1 23 22 15 13 14 6 8 2 11 8 24 16 6 11 4 0 0 0 0 ESSD 0 0 1 1 3 0 0 1 2 2 0 0 0 1 2 1 0 2 1 0 0 0 0 0 ESSL 0 0 0 1 2 1 0 0 3 2 0 1 1 1 4 2 0 0 0 0 3 0 0 0 ESSP 0 0 0 0 3 1 0 2 3 0 0 1 2 0 1 2 1 0 0 0 0 2 0 0 ESST 0 0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 ESSU 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ESSV 0 0 0 0 3 5 3 7 2 4 0 2 0 1 5 1 6 1 1 1 0 1 0 0 ESTA 0 0 0 0 2 2 1 6 1 0 1 2 1 1 2 1 1 3 1 0 1 0 0 0 ESTL 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 ESUP 0 0 0 1 0 0 0 0 0 1 0 0 0 0 1 0 0 1 0 0 0 0 0 0 ESUT 0 0 0 0 1 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 ;

param oAD: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 := #Monday