4
Web-based Intelligent Computational Argumentation based Conflict Resolution in Air Traffic Management Xiaoqing (Frank) Liu * Hojong Baik ** Ravi Santosh Arvapally * Rubal Wanchoo * * Department of Computer Science ** Department of Civil, Arch. and Environmental Engg. Missouri University of Science and Technology – Rolla, MO, USA {fliu, baikh, rsamb6, rwtyf}@mst.edu Abstract-Collaborative decision making (CDM) is a process of reaching consensus on a potential solution of an issue through the evaluation of the different possible alternatives. The web-based intelligent computational argumentation system allows concerned stakeholders to post their arguments on different alternatives, assign weights and priorities to the arguments and reach the most favorable alternative using fuzzy intelligent techniques over the Internet. Exchange of information among the stakeholders improves the collaboration and drives towards the collective intelligence. In this paper, we show how our tool facilitates resolution of conflicts in air traffic management. It could enhance the Ground Delay Program (GDP) and help the Air Traffic Control System Command Center (ATCSCC) to take a better decision depending on the argumentation of Air Route Traffic Control Centers (ARTCC) and Airlines. Keywords – Fuzzy Inference Engine, Fuzzy Association Memory Matrix, ATCSCC, ARTCC, GDP. I. INTRODUCTION The process of decision making involves a close analysis of various available alternative choices and identifying the best alternative that has the highest favorability. Any issue may involve many stakeholders who make strategic decisions and the decision thus made exhibits a consensus among those stakeholders. There is a lot of information exchange in the form of artifacts and arguments among the experts while designing the product. When these stakeholders are geographically dispersed, it is unmanageable to document the artifacts, discussions and arguments among them. The web-based intelligent computational argumentation system allows stakeholders to post their arguments on different alternatives of an issue and helps the group to reach an alternative which is favored by most of the stakeholders, and it facilitates resolution of conflicts by explicitly capturing rationale of stakeholders and reasoning about arguments. Every argument has a weight value which represents the degree of support or attack for that argument to an alternative or another argument and each argument influences the decision alternative. CDM is one of the most important aspects in any industry, and one such discipline is air traffic management where every decision made is in a high level strategic scenario. Here, the stakeholders are geographically dispersed across the country and the decisions made are mission critical. In this paper, we explain how our tool can be used in enhancing the Ground Delay Program (GDP) by demonstrating the technique through a developed and tested case study. In this paper, we use the words alternative and position synonymously. This paper is organized in the following manner: Section 2 reviews the previous work done in this area; Section 3 explains how the tool works and how it actually resolves the conflicts among the stakeholders. Section 4 explains how the web-based intelligent computational argumentation tool is applied in air traffic management, and section 5 presents the conclusion. II. RELATED WORK One of the important goals of Federal Aviation Administration (FAA) is to apply strategic initiatives to advocate anticipated demand-capacity imbalances at airports. When such an imbalance is expected at an airport, traffic managers apply ground delays to flights bound for the troubled airport commensurate with the delays they would receive in an airborne queue [3]. The current GDP rations the available arrival slots at the affected airport by scheduled arrival time of the flights with some adjustments to balance the equity between airlines. Current rationing rules do not take into account passenger flow efficiency in the rationing assignment tradeoffs [4]. Many scientists have examined different GDP rationing rules to achieve fairness among airlines; one of them was to distribute delays among the airlines equally. There are compressions and scheduling algorithms to overcome this problem. Air Traffic Control (ATC) specialists and CDM participating airlines use Flight Scheduled Monitor (FSM), developed by Metron Aviation Inc., to monitor and model Traffic flow Management. The FAA command center also known as ATCSCC, other FAA facilities, and the airlines use FSM to display Airport Demand List (ADL) information, monitor the airport-traffic situation, and collaborate on problems. Flight Schedule Monitor imports and displays ADL data, enabling all stakeholders to view airport demand and capacity, to list flights, to produce flight counts and statistics, and to color-code flights according to a variety of fields. All these developed models do not consider the problems associated with the airlines, they just pay attention to the fairness and efficiency of the model. There is very limited chance for airlines to argue with the FAA command center regarding the slot allocations in the present system, this is the main drawback. Web-Based intelligent argumentation system allows the airlines to argue on their issues. Several negotiation algorithms and other models are developed in a mediator-based argumentation system based on the extension of dung’s abstract model and they fall under 2010 10th Annual International Symposium on Applications and the Internet 978-0-7695-4107-5/10 $26.00 © 2010 IEEE DOI 10.1109/SAINT.2010.38 100 2010 10th Annual International Symposium on Applications and the Internet 978-0-7695-4107-5/10 $26.00 © 2010 IEEE DOI 10.1109/SAINT.2010.38 117 2010 10th Annual International Symposium on Applications and the Internet 978-0-7695-4107-5/10 $26.00 © 2010 IEEE DOI 10.1109/SAINT.2010.38 117 2010 10th Annual International Symposium on Applications and the Internet 978-0-7695-4107-5/10 $26.00 © 2010 IEEE DOI 10.1109/SAINT.2010.38 117 2010 10th Annual International Symposium on Applications and the Internet 978-0-7695-4107-5/10 $26.00 © 2010 IEEE DOI 10.1109/SAINT.2010.38 117

[IEEE 2010 10th IEEE/IPSJ International Symposium on Applications and the Internet (SAINT) - Seoul, Korea (South) (2010.07.19-2010.07.23)] 2010 10th IEEE/IPSJ International Symposium

  • Upload
    rubal

  • View
    212

  • Download
    0

Embed Size (px)

Citation preview

Page 1: [IEEE 2010 10th IEEE/IPSJ International Symposium on Applications and the Internet (SAINT) - Seoul, Korea (South) (2010.07.19-2010.07.23)] 2010 10th IEEE/IPSJ International Symposium

Web-based Intelligent Computational Argumentation based Conflict Resolution in Air Traffic Management

Xiaoqing (Frank) Liu* Hojong Baik** Ravi Santosh Arvapally* Rubal Wanchoo* *Department of Computer Science **Department of Civil, Arch. and Environmental Engg.

Missouri University of Science and Technology – Rolla, MO, USA {fliu, baikh, rsamb6, rwtyf}@mst.edu

Abstract-Collaborative decision making (CDM) is a process of reaching consensus on a potential solution of an issue through the evaluation of the different possible alternatives. The web-based intelligent computational argumentation system allows concerned stakeholders to post their arguments on different alternatives, assign weights and priorities to the arguments and reach the most favorable alternative using fuzzy intelligent techniques over the Internet. Exchange of information among the stakeholders improves the collaboration and drives towards the collective intelligence. In this paper, we show how our tool facilitates resolution of conflicts in air traffic management. It could enhance the Ground Delay Program (GDP) and help the Air Traffic Control System Command Center (ATCSCC) to take a better decision depending on the argumentation of Air Route Traffic Control Centers (ARTCC) and Airlines.

Keywords – Fuzzy Inference Engine, Fuzzy Association Memory Matrix, ATCSCC, ARTCC, GDP.

I. INTRODUCTION The process of decision making involves a close

analysis of various available alternative choices and identifying the best alternative that has the highest favorability. Any issue may involve many stakeholders who make strategic decisions and the decision thus made exhibits a consensus among those stakeholders. There is a lot of information exchange in the form of artifacts and arguments among the experts while designing the product. When these stakeholders are geographically dispersed, it is unmanageable to document the artifacts, discussions and arguments among them. The web-based intelligent computational argumentation system allows stakeholders to post their arguments on different alternatives of an issue and helps the group to reach an alternative which is favored by most of the stakeholders, and it facilitates resolution of conflicts by explicitly capturing rationale of stakeholders and reasoning about arguments. Every argument has a weight value which represents the degree of support or attack for that argument to an alternative or another argument and each argument influences the decision alternative. CDM is one of the most important aspects in any industry, and one such discipline is air traffic management where every decision made is in a high level strategic scenario. Here, the stakeholders are geographically dispersed across the country and the decisions made are mission critical. In this paper, we explain how our tool can be used in enhancing the Ground Delay Program (GDP) by demonstrating the technique through a developed and tested case study. In this paper,

we use the words alternative and position synonymously. This paper is organized in the following manner: Section 2 reviews the previous work done in this area; Section 3 explains how the tool works and how it actually resolves the conflicts among the stakeholders. Section 4 explains how the web-based intelligent computational argumentation tool is applied in air traffic management, and section 5 presents the conclusion.

II. RELATED WORK One of the important goals of Federal Aviation

Administration (FAA) is to apply strategic initiatives to advocate anticipated demand-capacity imbalances at airports. When such an imbalance is expected at an airport, traffic managers apply ground delays to flights bound for the troubled airport commensurate with the delays they would receive in an airborne queue [3]. The current GDP rations the available arrival slots at the affected airport by scheduled arrival time of the flights with some adjustments to balance the equity between airlines. Current rationing rules do not take into account passenger flow efficiency in the rationing assignment tradeoffs [4]. Many scientists have examined different GDP rationing rules to achieve fairness among airlines; one of them was to distribute delays among the airlines equally. There are compressions and scheduling algorithms to overcome this problem. Air Traffic Control (ATC) specialists and CDM participating airlines use Flight Scheduled Monitor (FSM), developed by Metron Aviation Inc., to monitor and model Traffic flow Management. The FAA command center also known as ATCSCC, other FAA facilities, and the airlines use FSM to display Airport Demand List (ADL) information, monitor the airport-traffic situation, and collaborate on problems. Flight Schedule Monitor imports and displays ADL data, enabling all stakeholders to view airport demand and capacity, to list flights, to produce flight counts and statistics, and to color-code flights according to a variety of fields. All these developed models do not consider the problems associated with the airlines, they just pay attention to the fairness and efficiency of the model. There is very limited chance for airlines to argue with the FAA command center regarding the slot allocations in the present system, this is the main drawback. Web-Based intelligent argumentation system allows the airlines to argue on their issues. Several negotiation algorithms and other models are developed in a mediator-based argumentation system based on the extension of dung’s abstract model and they fall under

2010 10th Annual International Symposium on Applications and the Internet

978-0-7695-4107-5/10 $26.00 © 2010 IEEE

DOI 10.1109/SAINT.2010.38

100

2010 10th Annual International Symposium on Applications and the Internet

978-0-7695-4107-5/10 $26.00 © 2010 IEEE

DOI 10.1109/SAINT.2010.38

117

2010 10th Annual International Symposium on Applications and the Internet

978-0-7695-4107-5/10 $26.00 © 2010 IEEE

DOI 10.1109/SAINT.2010.38

117

2010 10th Annual International Symposium on Applications and the Internet

978-0-7695-4107-5/10 $26.00 © 2010 IEEE

DOI 10.1109/SAINT.2010.38

117

2010 10th Annual International Symposium on Applications and the Internet

978-0-7695-4107-5/10 $26.00 © 2010 IEEE

DOI 10.1109/SAINT.2010.38

117

Page 2: [IEEE 2010 10th IEEE/IPSJ International Symposium on Applications and the Internet (SAINT) - Seoul, Korea (South) (2010.07.19-2010.07.23)] 2010 10th IEEE/IPSJ International Symposium

formal argumentation [6]. They may not bdecision environments with a lot of uncertAir Traffic Management.

III. WEB-BASED INTELLIGENT COMPUTATIONAL ARGUMENTATIONCONFLICT RESOLUTION SYSTEM

A. Background A Web-based intelligent computational

system has been developed to supportdecision making and it is based on the architecture [1]. All the services are residethat are accessible using a Web browser anaccessible from anywhere in the world viathis enables geographically scattered pecollaboratively by presenting their points ointelligent argumentation network, the arguis organized as a weighted directed graph [1root node denotes a position, nodes denotedcontain arguments and nodes denoted by oevidences supporting those arguments. Arcassociation between two nodes. The associtwo nodes can be one of the following support or indecisive. The weight assigargument is the strength of the argument anby a stakeholder. The weight value of an real number which can be in the range opositive weight value of an argument derelationship and a negative weight value relationship while zero denotes indecisivThe strength of the argument is viewed as alinguistic labels are used to represent

A set of twenty-five fuzzy inferendeveloped to assess indirect impact of ananother argument or decision alternativmanaged in a fuzzy associative memory fuzzy inference engine is developed basedinference rules. The inputs to the fuzzy inare the weight of an argument that has to bthe weight of argument which is right abargumentation tree. The output of the fuengine is the weight of the argument, wreduced by one level. In this way, the comargumentation network is reduced level by

P

A1 0.9

A2 -0.7

A3 0.2

E1 0.5

E2 0.8

Figure 1. Position Dialog Graph

Support Attack

Support

be effective in tainty, such as

N BASED

argumentation t collaborative

Client-Server ent on a server nd is therefore a the Web and eople to work of view. In the ument structure 1]. In Figure 1, d by rectangles oval shape are cs represent an iation between types: attack,

gned to each d it is assigned argument is a

of -1 and 1. A enotes support denotes attack e relationship. a fuzzy set and

the strength.

nce rules are n argument on ve. They are matrix [1]. A

d on the fuzzy ference engine be reduced and bove it in the uzzy inference which is now

mplexity of the level using the

fuzzy inference engine to the poiis directly associated with anfavorability factor of each altesumming up all current weights alternative with the maximum fmost preferred one [1]. The fuimplemented in an intelligent arcollaborative decision making. Fscreen shot of an argumentation tr

IV. APPLICATION IN AIR TRAMANAGEMENT A. Description

The Airline industry remainindustry and it is a significant economy and provides a service by other modes of transportatiohours in The United States, thereper hour in the sky. And the Nzones called ARTCC. ATCSCCits role is to manage the flow continental United States. ARTresponsible for controlling instruen route in a particular volumaltitudes. Ground Delay Progcontrol air traffic volume arouprojected traffic demand is eairport’s acceptance rate for a Demand exceeding the acceptaresult of the airport’s acceptancesome reason. The most common acceptance rate is adverse weathyear 1999 to 2006, on an averageyear were declared around the UGDP program is declared, thdownsize the number of flight opin order to achieve a balance bflights and the airport acceptanchas to make a reasonable solutiooperations in each airline while among all the airlines and sometibe happy with the number of fligh

Figure 2. Argum

nt where every argument n alternative. Then the rnative is computed by of these arguments. The

favorability factor is the uzzy inference engine is rgumentation system for igure 2 is an example of ree in the system.

AFFIC

ns a large and growing engine of the national that cannot be achieved

on [2]. During the peak e are about 5,000 flights NAS is divided in to 21

C is the command center; of air traffic within the CC is a facility that is

ument flight rules aircraft me of airspace at high gram is implemented to und airports where the xpected to exceed the lengthy period of time.

ance rate is normally a e rate being reduced for reason for a reduction in

her conditions. From the e, 960 GDP programs per United States. When the he ATCSCC needs to perations for each airline between the demand for e rate. So, the ATCSCC

on for reducing the flight maintaining the fairness

imes the airlines may not ht operations allocated to

mentation tree

101118118118118

Page 3: [IEEE 2010 10th IEEE/IPSJ International Symposium on Applications and the Internet (SAINT) - Seoul, Korea (South) (2010.07.19-2010.07.23)] 2010 10th IEEE/IPSJ International Symposium

them. Our tool allows the airlines to arguefor which they are not satisfied. This sectiocase study, which was developed and testeBy using this tool, ATCSCC can better uproblems of airlines and other stakeholdersarguments and take a better decision and it enhance the GDP planning process. B.Case Study

Let us suppose that due to incremconditions, a large-hub airport such as Cdecides to reduce its operational capacity.reduction will initiate GDP program, andflight operations needs to be discussed via camong stakeholders at ATCSCC. Reducingflight operation slots for each airline is an ito be dealt with. In our case study, the ATCthe issue and its possible alternatives. Othecan also post their alternatives if they alternative meets the criteria set by the ATCassume that the Chicago ORD airport hoperations per hour. Because of GDP progoperations have to be reduced to 45 – 60 hour. The length of GDP affected period is one hour and it is also assumed that the occurs during the day time. The Airlines1 that airport and the Airlines3 operates mointernational flights than domestic flights duaffected hour. The following TABLE sairlines involved in this case study along woperations during that GDP affected hour. TABLE I. FLIGHT OPERATIONS OF AIRLINE

Airlines Flight Operations / hour Airlines1 40 operations / hour Airlines2 24 operations / hour Airlines3 36 operations / hour

We have five stakeholders involved inmaking group. Every stakeholder is given asystem, which must be a value ranging fropriority is used to assess the strength of anit influences the favorability factor of an aThe following TABLE shows priority of eain this case study. TABLE II. PRIORITIES OF THE STAKEH

There are five alternatives for the gwhich the first two alternatives are proATCSCC, and the remaining three alprovided by the airlines. The first two alterthe equity, and all the airlines are given eqoperational slots. Each and every alternacriteria set by the ATCSCC for an alternanumber of flight operations has to be betweThe TABLE III shows, the total numballocated to each airlines according to each

e on the issues on deals with a ed on our tool. understand the s through their could possibly

mental weather Chicago ORD This capacity

d reducing the conference-call g the number of important issue CSCC will post er stakeholders think that the CSCC. Let us has 100 flight gram, the flight

operations per assumed to be GDP program has its hub in

ore number of uring that GDP shows all the

with their flight

ES

n the decision a priority in the m 0 to 1. This

n argument and alternative [5].

ach stakeholder

HOLDERS

given issue, in ovided by the lternatives are rnatives follow qual number of ative obeys the ative; the total een 45 and 60. ber of flights alternative.

TABLE III. FLIGHT SLOT ALLOCATI

C. The Argumentation Framewor

This section explains how thargumentation tool works for AInitially, the Air Traffic Control Sidentifies the issues and its postakeholders can now post thalternatives listed by ATCSCCarguments against an alternativethey can even post supportinarguments. An argument can suargument also. Once the argusystem computes the favorabialternatives. Figure2 shows the ar

Web-Based intelligent argum

based framework for argumentatrees. The argumentation tree grpost their arguments under theargument is posted, the stakeholhis/her name, weight of the reassess the weight of an areassessment technique discussedmechanism specified in [1], the aa single level. Finally, the weightis used to compute the favalternative. Figure 3 and Figure trees of alternative 1 and alternaargumentation trees for alternatialternative 5 are similar to alterna

A

Iss

Flight

Malt

Evidences

Arguments & Weights

Attack/Support

Figure 3. Argumenta

ION ALTERNATIVES

rk he web-based intelligent

Air Traffic Management. System Command Center ossible alternatives. The heir arguments on the C. They can post their e or in support of it and ng evidences to their upport or attack another

umentation is done, the ility factor of all the rgumentation framework.

mentation tool is a logic-ation based on argument rows as the stakeholders e alternatives. When an lder is supposed to enter argument. Initially, we rgument using priority d in [5]. Then using the arguments are reduced to ted summation technique

vorability factor of an 4 are the argumentation

ative 2 respectively. The ive 3, alternative 4 and

ative 1 and alternative 2.

Air Traffic Stakeholders

Conflicts

sues of flight slot

t slot allocation alternatives

ost favorableternative

ation Framework

102119119119119

Page 4: [IEEE 2010 10th IEEE/IPSJ International Symposium on Applications and the Internet (SAINT) - Seoul, Korea (South) (2010.07.19-2010.07.23)] 2010 10th IEEE/IPSJ International Symposium

The detailed arguments of the boxes are a A1-This alternative has minimum numoperations among all the alternatives. A1.1-It satisfies the range of 45-60 flight hour as suggested. A2-There is no equity problem in thFairness is maintained among the airlines. A3-This alternative operates 50 flight operaIt is the best one among all the alternatives. A4-It is difficult to cut down 50% of flightbetter if is 40% cut down, still the equity is A4.1-This idea would be really great, I cfinancial loss to a great extent. A4.2-Passenger delay could be reduced. A4.3-The sector workload will be relativelyA5-Workload in sectors is relatively bealternative. A6-I have my hub in this airport, I need lemy flights. 50% really affects my economy.A6.1-Customer satisfaction and reputation goes down with this alternative. A6.2-I do not have any flight operation slowith you. A6.2.1-I am running short of flight operatinot in a position to exchange slots. B1-This alternative is better than alternativenumber of flight operations. B2-This alternative has 55 flight operations a good alternative. B3-I have many international flights during th

Figure 3. Argumentation tree of Altern

Figure 4. Argumentation tree of Altern

as follows: mber of flight

operations per

his alternative.

ations per hour. ts. It would be maintained.

can reduce my

y more. etter with this

ss cut down of . of the airlines

ots to exchange

ion slots. I am

e1. It has more

per hour. It is

his time. This

alternative doesn’t work with me. B3.1-We can exchange a flight opB4-This alternative has more sectB5-I have hub in this airport, so Ioperation slots for me. D. The Favorability factor

Figure 5 shows the favorabilalternatives. Alternative4 is the mwith a favorability factor of 0.907

V. CONCLUSION

This paper discusses how to uargumentation to facilitate resoltraffic management. Web-based argumentation based conflictinformation exchange among thegeographically distributed locatiotraffic management, ATCSCC cviewpoints and preferences of airdecision. It has the advanttransparency by allowing stakehtheir views and preferences by palternatives and arguments and iwork.

ACKNOWLEDGMThis research is funded by CInfrastructure and Safety - NationalCenter at Missouri University of Scie REFERENCES [1] Liu, X.F., S. Raorane, and M. LeCollaborative System for Engineering Dedesign and manufacturing methodologieS.K. Ong, Andrew Y.C. Nee, Chris McMpp. 37-58. [2] Duke and Torres, “Multifactor prtransportation industry,” vol. 128, Monthpp. 32-45. [3] R. Hoffman, M. Ball, and A. MukDistance with equity guarantees: A neprogram planning and control,” the 7th ASpain, 2007. [4] B. Manley and L. Sherry, “The Imp(GDP) Rationing Rules on Passenger Proceedings of the IEEE Integrated ComSurveillance Conference, May 2008. [5] Liu, X.F., E. Khudkhudia, L. Wen, V. Intelligent Computational ArgumentatCollaborative Software Development DeIntelligence Applications for ImproDevelopment, Farid Meziane and SuniGlobal, 2009, pp. 167-180. [6] Barbini, P., Wu, Y., and Caminadargument based discussion,” In ProceedConference on Autonomous Agents and MBudapest, Hungary, May 10 - 15, 2009.

native 1

native 2

Figure 5. Resu

. peration slot. tor work load relatively. expect more flight

lity factor of all the five most favorable alternative 7.

use web-based intelligent lution of conflicts in air intelligent computational t resolution improves e stakeholders who are in ons. When applied in Air an better understand the rlines and make a proper ages of fairness and

holders explicitly express posting issues, solutions, it improves collaboration

MENT enter for Transportation, l University Transportation ence & Technology - Rolla.

eu, “A Web-based Intelligent sign,” in Collaborative product es and applications, W.D. Li,

Mahon, London: Springer, 2007,

roductivity change in the air hly Labor Review, March 2005,

kherjee, “Ration-By- ew approach to ground delay

ATM R&D seminar, Barcelona,

pact of Ground Delay Program and Airline Equity,” In the

mmunications, Navigation and

Sajja, and M. Leu, “An ion System for Supporting ecision Making,” in Artificial oved Software Engineering il Vadera, Hershey, PA: IGI

da, M.,”An implementation of dings of the 8th international Multiagent Systems - Volume 2

ult

103120120120120