Optimum Fleet Utilization under
Congestion Management at NY LGA
George L. Donohue, Ph.D.
Professor Systems Engineering and Operations Research
Director of the Center for Air Transportation Systems Research
Volgenau School of Information Technology and Engineering
NEXTOR Wye River Conference June 7, 2007
CENTER FOR AIR TRANSPORTATION SYSTEMS RESEARCH
© George L. Donohue 2007
Credits
Research team at GMU contributing to
these insights:
•
•
•
•
•
Dr. Loan Le, Ph.D. (2006)
Dr. Karla Hoffman, Prof. SEOR, CATSR
Danyi Wang, Ph.D. Candidate
Ning Xie, Ph.D. Candidate
Dr. C.H. Chen, Prof. SEOR, CATSR
RAND Corp.:
• Dr. Russell Shaver, Senior Research Fellow
2
CATSR
Outline
•
•
•
•
3
Motivation for the Study
NY LaGuardia Data
Approach
Results from Schedule Optimization and
Delay Simulation Study
CATSR
Annual Passenger Enplanements Predicted to be
Lost: FAA Forecast to 2025
CATSR
Annual Projected Enplanements Foregone
Because of Airport Capacity Constraints
ORD
20
JFK
2025
2015
15
2005
EWR
10
LGA
Airports
STL
TPA
SLC
SEA
SFO
PIT
SAN
PHL
PHX
PDX
ORD
MIA
MSP
MEM
MCO
MDW
LAX
LGA
LAS
IAH
JFK
IAD
FLL
HNL
DTW
EWR
DEN
DFW
DCA
CLT
CVG
BWI
CLE
ATL
5
0
4
Optimistic: All
Planned Airport
Improvements
Occur
All available landing slots fully
utilized regardless of congestion
25
BOS
Annual Enplanements Lost (Millions)
30
FAA 2006 TAF & 2004 Benchmark
Estimated Annual Cost to US (Lost Consumer Surplus,
2005$) due to Expected Airport Capacity Limitations
Annual Cost to US Economy from
Foregone Enplanements ($B)
$16
$12
FAA Assumptions on
Growth in Airport
Operations and
Passenger
E l
t
$10
Assumes:
$200/segment ticket
Price Elasticity = -1
$14
CY2015, All A/P
Improvements
CY2015, No A/P
Improvements
$8
CY2025, All A/P
Improvements
$6
$4
CY2025, No A/P
Improvements
$2
$0
100%
5
95%
90%
85%
Usable NAS Capacity (%)
80%
Shaver 2007
CATSR
Severe Congestion at HDR Airports:
A 40-year-old Reality
CATSR
Timeline recap of congestion management measures
HDR at EWR, LGA, JFK,
DCA, ORD
Perimeter rule at LGA,
DCA
1969
- Limited #IFR slots
during specific time
periods
- Negotiation-based
allocation
6
early 1970s
Deregulation
1978
Removal of HDR
at EWR
Introduction of Huband-Spoke Network
System
Slot
ownership
AIR-21
1985
4.2000
Use-it-orlose-it rule
based on
80% usage
Exempt from
HDR at LGA,
JFK, ORD
certain flights
to address
competition
and small
market access
How is the Public Best Served?
Top 20 Worst Airports in the US (45-PTD)
Year
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
7
2004
Airports
ORD
EWR
LGA
PHL
ATL
MIA
FLL
MCO
DFW
LAS
BOS
SFO
IAD
JFK
CLE
SEA
TPA
STL
PDX
BWI
Prob. Of
PaxDelay
>45 min
14%
14%
13%
12%
11%
9%
9%
9%
9%
9%
9%
9%
9%
9%
9%
8%
8%
8%
8%
8%
2005
Prob. Of
PaxDelay
Airports >45 min
18%
EWR
17%
LGA
14%
ATL
13%
PHL
13%
BOS
12%
ORD
12%
FLL
12%
JFK
11%
MIA
11%
SFO
10%
SEA
10%
IAD
10%
TPA
10%
MCO
9%
BWI
9%
PIT
9%
PDX
9%
DTW
9%
LAS
9%
DCA
2006
Airports
ORD
EWR
LGA
PHL
JFK
IAD
MIA
ATL
MDW
DTW
DFW
BOS
DEN
CLT
IAH
CLE
PIT
DCA
MEM
SFO
Prob. Of
PaxDelay
>45 min
17%
16%
15%
15%
14%
12%
12%
12%
12%
12%
12%
11%
11%
10%
10%
10%
10%
10%
10%
10%
Average of 2004 to 2006
Airports
EWR
LGA
ORD
PHL
ATL
JFK
BOS
MIA
FLL
IAD
DFW
SFO
DTW
MCO
LAS
CLE
PIT
SEA
MDW
DCA
Prob. Of
PaxDelay >45
min
16%
15%
15%
13%
12%
11%
11%
11%
10%
10%
10%
10%
9%
9%
9%
9%
9%
9%
9%
9%
D. Wang, GMU PhD. In Progress
CATSR
EWR a NYNJ Airport with No Slot Controls:
Market Acceptable Transportation Predictability (2006)
Delay per Flight (minutes)
EWR
80
60
40
20
35 Airport BN Model
0
Summer 2006 Data
-20
-40
(N. Xie GMU 2007)
0
5
10
15
Time of Day (hour)
45 minute Passenger Trip
Delay (45-PTD) Metric
that Includes Flight Load
Factors, Cancellations &
Missed Connections
(D. Wang GMU 2007)
8
20
CATSR
Calculated Capacity (Today) and Actual Throughput
Optimum Rate
CATSR
80
Calculated Capacity - Today
Facility Reported Rate - EWR
(arrivals, departures per hr)
Arrivals per Hour
42, 42
60
EWR :
DoT/FAA
40
Infrequent
2004 Capacity
Benchmark Report
Most Frequent
20
Each symbol represents actual
traffic during a single hour
0
0
20
40
60
80
Departures per Hour
Marginal Rate
IFR Rate
80
60
Arrivals per Hour
Arrivals per Hour
80
40, 40
40
20
60
33, 33
40
20
0
0
0
9
Operational Rates
Inside this Envelope
Satisfy FAA Safety
Standards (i.e.
establish Max.
Number of Arrival
and Departure Slots)
20
40
60
Departures per Hour
80
0
20
40
60
Departures per Hour
80
Are Airlines Making Optimal use of these
Operations? EWR Fleet Mixture
Monthly Averaged
T-100 Segment Data
10
CATSR
New York LaGuardia Airport: Case Study
Data (2005):
• Throughput:
•
•
•
•
11
404,853 flights/yr
Average flight delay:
38 min
Revenue passengers:
26,671,787
Average aircraft size:
96 passenger
Average inter-city fare:
$133
CATSR
NYNJ Airport with Current Slot Controls:
LGA 2004 - 2006
Calculated Capacity (Today) and Actual Throughput
LGA
Optimum Rate
80
60
Calculated Capacity - Today
60
39,39
Facility Reported Rate - LGA
(arrivals, departures per hr)
40
Arrivals per Hour
Delay per Flight (minutes)
CATSR
20
0
40
Infrequent
20
-20
-40
Most Frequent
Each symbol represents actual traffic
during a single hour
0
0
5
10
15
20
0
Time of Day (hour)
20
40
60
Departures per Hour
Marginal Rate
IFR Rate
60
60
37,37
Arrivals per Hour
Arrivals per Hour
37,37
40
20
0
12
40
20
0
0
20
40
Departures per Hour
60
0
20
40
Departures per Hour
60
Current Government Rules at LGA Also
Lead to Poor Use of Runway Resources
• Inefficient use of
resources
Airports win
Airlines win
(High Load
Factor/Large
Aircraft)
Airports lose
Airlines lose
(Low load
factor/Small
Aircraft)
13
CATSR
Why do the Airlines Schedule beyond the Maximum Safe
RW Capacity with Flights that Loose Revenue?
• There is no government regulation to Limit schedules for
Safety or Compensate passengers for Excessive Delays
and Volume related Flight Cancellations
• These were errors in the 1978 Deregulation Act
• Congress creates New Slots
• Passenger surveys indicate that Frequency and Price are
the most (Only?) desirable characteristics of a flight
• Passengers are not Told of Consequences of published
schedule to travel Predictability
• If any One airline decided to offer Feasible Schedules,
their competition might offer more frequency to capture
market share (No Good Deed goes Unpunished!)
• Thus, still producing delays and cancellations for All
• In Game Theory, this is called the Prisoner’s Dilemma
14
CATSR
A Natural Question? Is There an Optimal
Allocation of Scarce Runway Resources?
• What would happen if schedules at major airports were
capped by predictable runway capacity and allocated by a
market mechanism?
• What markets would be served?
• How would airline schedules change?
– Frequency
– Equipment (#seats per aircraft)
• How would passenger demand change?
– At airport
– On routes
• How would airfares change?
– What would happen to airline profit margins?
• How would airport and network delays be altered?
15
CATSR
Modeling Approach and Assumptions
• A Benevolent Monopolistic Airline (e.g. Port
Authority of NY&NJ) has the ability to
Determine and Set an Optimum Schedule to:
• Operate at Competitive Profit Margins
• Maximize Passenger Throughput
• Ensure an Airline Operating Profit (Max, 90%,80%)
• All Current Domestic Origin and Destination
Markets are Considered
• 67 Scheduled Daily Service Markets
• Current Market Price Elasticity Remains
Constant
• BTS T-100 segment data
16
CATSR
NY LGA Has 67 Daily Markets
17
CATSR
Airline Competitive Scheduling:
Modeling Framework
Auction 64
Ops Slots/Hr
Demand-Price Elasticity
$
ASPM, BTS
databases
S1
S2
D
#
Network Flow
Optimization Problem
Flight schedules
Fleet mix
Average fare
Flight delays
Delay Network
Simulation
18
(Le, 2006)
CATSR
Research Results:
Detailed Data at 90% of Profit Optimality
CATSR
Estimated Effect of Slot Controls at LGA Using Market Mech
Market Change
# flts/day
Max Enpl/day
Avg AC Seat Cap
Avg Fare
Avg Flt Delay
100%
Percent Change from Current Policy
80%
60%
40%
LGA IMC
Capacity
Aircraft Size
20%
Airfare
0%
-20%
0
BL
1
U
2
10
3
94
58
76
76
49
10
Max enpl/day
-40%
Markets served
-60%
Frequency of Service
-80%
Delays
-100%
-120%
Arrival Rate per 15min Time Slot
19
58
90% Profit Max - 64 Ops/Hr compromise:
Frequency and Delay profile by time of day
20
CATSR
Optimized Schedule Frequency and Aircraft
Gauge by Market (Opt/Current)
Optimized Frequency and Aircraft Gauge
3.5
Optimum to Current Ratio
3.0
2.5
Freq. A/C Gauge
0.8
1.2
Mean
0.9
1.1
Median
1
1
Mode
0.23
0.50
SD
Flight Frequency
Aircraft Gauge
2.0
1.5
1.0
0.5
0.0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63
> 2/Hr
21
Frequency Rank
CATSR
Model Estimate of Aircraft Gauge Change
CATSR
Estimate of Aircraft Up-Gauging
300
Current Fleet Allocation
90% Optimum Fleet Allocation
Number of Daily Flights
250
200
Aircraft Size Range
19 to 37
44 to 50
69 to 77
88 to 110
117 to 133
138 to 158
166 to 181
194 to 225
Total
150
100
50
0
19 to 37 44 to 50 69 to 77
22
88 to
110
117 to
133
138 to
158
Aircraft Seating Capacity
166 to
181
194 to
225
Current
# Flights
64
259
90
192
121
235
49
0
1010
90% Opt
# Flights
66
54
90
118
86
252
72
68
806
%
Change
3%
-79%
0%
-39%
-29%
7%
47%
-20%
Unprofitable daily markets at LGA
CATSR
• Three markets (13 Flights) that are not
profitable to operate on a daily basis are
identified to be:
•
•
•
Lebanon-Hanover, NH (LEB),
Roanoke Municipal, VA (ROA),
Knoxville, TN (TYS).
Runway Cap. Market seats/AC Fare Passengers RPM Yield Flights/day
unconstrained LEB
19
$153
50
$0.72
6
10,9,8,7
ROA
37
$186
77
$0.46
5
6,5,4
TYS
50
$125
85
$0.19
2
23
Research Results – Win Win
CATSR
Airlines adapt with aircraft size and frequency to congestion
constraint:
Positive impacts on passengers, airports, airlines, and ATC
•Airlines
• Reduced frequency with
larger aircraft
• Most Markets Retained
• More Profitable (90% of
Optimum)
•Passengers
24
• Markets served: Little
change
• Airfares no change
• Improved Predictability
• Airports
• Increased passenger
throughput
• Reduced delays (70%)
•Air Traffic Control
• Reduced delays
– Demand within capacity
– Reduced Prob. SRO
Conclusions and Recommendations
• Airport Congestion Management will be
Required to Accommodate Projected passenger
growth rates
• Market Based Approaches May be able to
Approximate Optimum Allocation of Scarce
Runway Availability Resources
• Metropolitan “Metroplex Operation” should be
Investigated to Better Understand Airport
Synchronization Possibilities under Congestion
Management Measures
25
CATSR
Opinion
• FAA Owns slots Because:
• FAA computes Max Number of Safe Arrival and
Departure Combinations as a Function of:
– Airport Runway Configuration,
– Separation Technology
– Designated Level of Safety
• Daily GDP control (and Acceptance by Airlines) is an
implicit exercise of this ownership
• Slot Exemptions (or total lack of control) are an
implicit reduction of the FAA’s stated safety
standards
• Standards should be either changed or enforced
26
CATSR
Center for Air Transportation System Research
Publications and Information
• Loan Thanh Le, “Demand Management at Congested
Airports: How far are we from Utopia?”, Ph. D.
dissertation August 2006.
• http://catsr.ite.gmu.edu
– Other Useful Web Sites
• http://mytravelrights.com
• http://gao.gov
• http://www.airconsumer.ost.dot.gov
27
CATSR
CATSR
BACKUP Material
28
DCA a Conservatively Scheduled High
Demand Slot Controlled Airport
CATSR
Calculated Capacity (Today) and Actual Throughput
Optimum Rate
80
60
Calculated Capacity - Today
60
Facility Reported Rate - DCA
(arrivals, departures per hr)
36, 36
40
Arrivals per Hour
Delay per Flight (minutes)
DCA
20
0
40
Infrequent
20
Most Frequent
-20
-40
Each symbol represents actual traffic
during a single hour
0
5
10
15
20
0
Time of Day (hour)
0
20
40
60
Departures per Hour
Marginal Rate
IFR Rate
60
60
Arrivals per Hour
Arrivals per Hour
30, 30
40
20
0
29
40
24, 24
20
0
0
20
40
Departures per Hour
60
0
20
40
Departures per Hour
60
Congestion Management could Shift Hubbing
Passengers to other Large Airports
Connecting
Airport
Passengers
%
Chicago O'Hare
59
Newark NJ
32
NY LaGuardia
8
NY JFK
40
Philadelphia
38
Atlanta
66
Boston
15
Miami
55
Washington Dulles
53
Dallas/Fort Worth
60
30
FAA 2006 NPIAS
CATSR
LGA High Frequency Flights:
Current and 90% of Optimum
Boston Logan
Washington DC Reagen Nat
Chicago O'Hare
Atlanta Hartsfield
Fort Lauderdale Fl
Raueigh/Durham NC
Detroit Mi
Charlotte NC
Columbus OH
Dallas Ft Worth
31
Market Daily
Freq
BOS
73
DCA
69
ORD
62
ATL
48
FLL
43
RDU
37
DTW
32
CLT
32
CMH
26
DFW
26
A/C
seats
106
108
138
156
157
46
122
102
46
148
Model Model Normalized
Freq
seats
Freq
60
208
0.8
68
131
1.0
56
139
0.9
32
145
0.7
26
181
0.6
22
95
0.6
22
175
0.7
30
97
0.9
22
102
0.8
26
146
1.0
CATSR
Rank
Seats
2.0
1.2
1.0
0.9
1.2
2.1
1.4
1.0
2.2
1.0
1
2
3
4
5
6
7
8
9
10
Model preserves Heterogeneous Aircraft Mix:
But Reduces Frequency and Up-Gauges some Markets
Model Schedule at 90% Optimum
250
2005 Schedule
Model Sched @ 90% Opt
Average Seats per Flight
200
150
100
50
0
0
10
20
30
40
50
Daily Flight Frequency
32
60
70
80
CATSR
Network Delays Driven by Uncoordinated and
Over-Scheduled Flights: e.g. LGA, EWR, JFK
CATSR
Airline Demand Driven by Market
Access/Competition/Profitability
Concerns
Runway Capacity is Set by Aircraft
Safety Separation Standards
1 Hr
33
•High utilization rates (>80%)
increase delays exponentially
•Delays at major airports Impact
the Entire Air Transportation
Network
JFK Fleet Mixture
34
CATSR
Sub-problem: Modeling a single market
CATSR
Build timeline network: complete schedule of all possible flights, fleets
Estimate arc costs
Estimate node revenues: for each 15-min arrival time window
Available data: 10% ticket price sample by quarter
35
ORDÆLGA segment, 100%
1000
900
800
700
1000
900
800
700
600
500
400
600
500
400
fare ($)
fare ($)
ORD-LGA segment, 10%
300
300
200
100
200
100
5000
10000
#passengers
15000
20000
200K
400K
#passengers
600K
Sub-problem: Modeling a single market
CATSR
Build timeline network: complete schedule of all possible flights, fleets
Estimate arc costs
Estimate node revenues: for each 15-min arrival time window
From demand curves to revenue curves
40K
peak to
off-peak
Revenues ($)
35K
30K
25K
20K
15K
10K
5K
36
General solution approach
Airline Profit
maximizing
ATL
schedules
fleets
CATSR
Airport’s
enplanement
opportunities
maximizing
BOS
67
nonstop
daily
domestic
markets
…
LGA
ORD
TPA
…
37
Multi-commodity
network
subproblems
marginal value
of allowing
an extra flight
(dual prices)
Set packing
master problem
Congestion management by
better scheduling
Schedules exceeding airport optimum rates…
38
CATSR
Congestion management by
better scheduling
…lead to excessive queuing delay
39
CATSR
Excess of demand and severe congestion at NY
CATSR
area airports: a 40-year old reality
HDR
1969
40
1985
Grandfather rights
- LGA: 75 operations/h
ADR:
Airport
Departure
Rate
- AIR21
added
additional
exemption slots
AAR:
Airport
Arrival
Rateto small airports
for
70-seat
or less
aircraft
Æ 600 requests = 50% more flights
Facility-reported capacity: average of (ADR+AAR)
4.2000
over 2000-2005 period
Excess of demand and severe congestion at NY
CATSR
area airports: a 40-year old reality
HDR
1969
41
1985
Grandfather rights
Removal of HDR
at ORD
6.2002
AIR21
4.2000
1.2001
Lottery
Removal of HDR at
LGA and JFK
What’s next?
1.2007
Summary of European Passenger Bill of Rights http://news.bbc.co.uk/1/hi/business/4267095.stm
•
Overbooked Flights
•
•
•
•
•
•
•
Offered a refund of your ticket, along with a free flight back to your initial point of
departure, when relevant. Or, alternative transport to your final destination.
Rights to meals, refreshments, hotel accommodation if necessary, even free e-mails or
telephone calls.
– Airlines can only offer you a refund in the form of travel vouchers if you agree in
writing
Refunds may also be paid in cash, by bank transfer or cheque
If the reason for your flight's cancellation is "within the airline's control", it must pay
compensation.
Compensation for cancellations must be paid within seven days.
Delayed Flights
•
42
Passengers can now get roughly double the existing compensation if they are bumped off a
flight.
– Compensation must be paid immediately.
– These passengers must also be offered the choice of a refund, a flight back to their
original point of departure, or an alternative flight to continue their journey.
May also have rights to meals, refreshments, hotel accommodation if necessary even free emails, faxes or telephone calls.
Cancelled Flights
•
•
CATSR
•
Airline may be obliged to supply meals and refreshments, along with accommodation if an
overnight stay is required.
If the delay is for five hours or more, passengers are also entitled to a refund of their ticket
with a free flight back to your initial point of departure if this is relevant.
Air Transportation System (ATS) is a
Network with 6 Interacting Layers
•The ATS is a Public - Private Partnership with
conflicting objective functions:
•Public – Commerce and safety; interest groups
•Private – Profit maximization
Passenger/Cargo Layer (Delays, Cancellations)
Airline Layer (Routes, Schedules, A/C size)
TSA/FAA Layer (ATC Radar, Radios, Ctr’s, Unions)
Weather Layer (Thunderstorms, Ice Storms)
Physical Layer (i.e. Cities, Airports, Demographics)
Government Regulatory Control Layer
43
CATSR
Air Transportation is Characterized as a
Complex Adaptive System (CAS)
Fleet
Attributes
Fleet
Costs
Trips flown
by fleet
CATSR
Fleet
Revenue
Effect on
GDP
Passenger
Delays
Delays
Airport
Capacity
Enroute
Capacity
Flight Delays
&
Cancellations Offered
Schedule
Inconven
ience
Reference
demand
44
Airline
Profits
Effective
Price
Flights by
Fleet
Market
Clearing
Ticket
price
Aircraft
Fleets
Active
fleet
Effective
price by
length of trip
Baseline
Demand
Bengi Mezhepoglu, PhD
in progress
Research problems and findings
Research problem 1:
Are current rules of slot allocation the main causes of the congestion
problem?
Answer:
Yes, grand-father rights, weight-based landing fees, slot exemptions
Research problem 2:
Impacts on congestion, enplanement opportunities, markets served,
aircraft size, flight demands ?
45
CATSR
Airline response model
CATSR
Model a single benevolent airline
Model the interaction of demand and supply through price
– Price elasticity of demand determine demand at each price point
– Each supply curve corresponds to a fleet mix profile
– Different supply levels result in different equilibriums
E2
E3
revenue ($)
fare ($)
E1
E1
E2
E3
46
quantity (#seats)
quantity (#seats)