A member of
FAA Center of Excellence
Air Transportation System (ATS)
Analysis:
Dynamic System Model of the ATS
Lance Sherry (Ph.D.)
Bengi Mezhepoglu (Ph.D. Candidate)
12/2005
CENTER FOR AIR TRANSPORTATION
SYSTEMS RESEARCH
Education
Analysis
Ed
ti
A l i and
d Research
R
h for
f the
th Next
N t
1
Acknowledgements
• Dave Knorr, Anne Yablonski (FAA)
• Dan Goldner (Ventana Systems)
• Mike Wambsgnass, Terry Thompson (Metron
Aviation)
• JP Clarke (Ga Tech), Mike Ball (UMd), Mark
Hansen (UC-B)
• Sherry Borener, Peggy Gervasi (JPDO)
• George Donohue, John Shortle, C.H. Chen,
Sasha Klein, Jonathan Drexler, Danyi Wang
(GMU)
2
Summary
• U.S. Domestic Air Transportation System
(ATS):
– amazing success story over last century
– significant contributor to the U.S. economy
• ATS is faced with (9) major challenges
• Use Dynamic System Model of ATS to
understand:
– cause-effect of challenges ?
– can system “self-heal” ?
3
Summary
•
Results of analysis:
1. ATS exhibits Time Constants and Equilibrium Points
that should be taken into account in strategic plans
•
•
Key metrics: (1) effective capacity, (2) demand
75% of capacity rule
2. Sustainability in a future ATS (e.g. NGATS) can be
attained by “designing” mechanisms to:
a)
b)
c)
balance capacity and demand
signal need for capacity enhancement
incentivize innovation
4
Organization
1.
2.
3.
4.
Success & Challenges
Air Transportation System Model
Results of Analysis
Opportunities
5
A member of
FAA Center of Excellence
1. Successes, Challenges
CENTER FOR AIR TRANSPORTATION
SYSTEMS RESEARCH
Education
Analysis
Ed
ti
A l i and
d Research
R
h for
f the
th Next
N t
6
Air Transportation Faces Challenges
7
Success Story – Capacity
ATS Capacity
800000
700000
600000
500000
400000
300000
200000
100000
0
Years
Domestic Enplanements
Domestic RPMs (Millions)
Domestic ASMs (Millions)
8
Success Story - Airfares
Domestic Real Yield (1978 Cents)
50
Jets
45
Widebodies,
Radar
40
1978 Cents
35
Hub-&-Spoke
30
25
Yield Mgmt
20
15
TFM, Internet,
LCCs
10
5
0
1920
1940
1960
1980
2000
2020
Year
Source: ATA (2005)
Yield = Before Tax Revenue per Seat-Mile
9
Challenges ?
1.
Airline Profit/Loss Cycles
–
2.
Equal Access
–
–
3.
wasted resources and low predictability
Environmental issues (emmissions, noise)
Airport & Airspace Trust Fund eroding
Airport & Airspace innovation cycle is dormant
ATC innovation cycle is dormant
–
9.
delays and cancellations
Congestion
•
5.
6.
7.
8.
Airlines consolidating service to high socio-economic metro areas
Eroding access from small communities (Essential Air Service subsidies)
Passenger Experience
–
4.
Finding economic equilibrium?
Modernization efforts effectively stalled
ATC labor issues (salary, staffing)
10
Challenges ? – Airline Finances
20,000,000
140,000,000
15,000,000
120,000,000
10,000,000
100,000,000
5,000,000
80,000,000
0
60,000,000
-5,000,000
40,000,000
-10,000,000
Annual Revenues
Cum Net Proft/Loss & O perating
Proft ($)
Airline Finances
20,000,000
-15,000,000
-20,000,000
0
Cum Net Profit/Loss
Source: ATA (2005)
Annual Operating Profit ($K)
Operating Revenues ($K)
Symptom, not a cause
11
Challenges – Passenger Experience?
800
450
Average = 1.5%
FLL-ORD, PHL-CLT
700
600
Number of Routes per Day
Gate-to-Gate Velocity (knots)
400
Velocity = 54.874Distance
350
300
250
2
R = 0.9068
200
150
100
0.2859
S A N - B WI
S LC - B WI
LA S - H O U
50
500
0
0 - 1%
1 - 2%
2 - 3%
3 - 4%
4 - 5%
5 - 6%
6 - 7%
7 - 8%
8 - 9%
Percentage of Flights Cancelled on Route
400
300
200
100
0
Origin to Destination Distance (nm)
Data Prepared By: Danyi Wang
12
Challenges – Passenger Experience
Passenger Delay per Enplanement (mins)
120
100
JFK-ORD
JFK-ORD
PHL-EWR
JFK-PHL
80
EWR-PHL
EWR-PIT
JFK-PIT
60
PHL-JFK
STL-PIT
PHL-STL
EWR-BWI
JFK-
LAS-M CO
M CO-LAS
ORD-JFK
40
LA S-DCA
DCA -LA S
20
0
0
500
1000
1500
2000
2500
Route Distance
13
Challenges ? - Congestion
•
Number of Flights
250
Unimpeded Actual Block Time
110 mins, (σ = 11 mins)
200
•
Scheduled Block Time
124 mins, (σ = 7 mins)
Median Actual Block Time
133 mins
150
Mean Actual Block Time
140 mins, (σ = 43mins)
100
•
•
Properties of system adds average 30
mins to a 110 minute flight (28%)
Properties of system results in
significant variability that makes
planning difficult
– Actual µ/σ = 3.25
– Schedule µ/σ = 20
Congestion fuel costs to airlines = 23% of revenue from flight
Adjusting schedule to mean actual
block time costs airline 12% in aircraft
utilization
50
0
100
90
200
150
190
290
>390
Atlanta – O’ Hare Actual Block Time (min)
Schaar, Drexler, Sherry (2005)
14
Challenges ? – A&ATF
Airport & Airway Trust Fund
$18,000
$16,000
$14,000
$M
$12,000
$10,000
$8,000
$6,000
$4,000
$2,000
-
Expenses
Source: ATA (2005)
Income
Balance
Ticket Tax Income
Shift to RJs, results in more ATC operations, A&TFless revenue
15
Challenges ? – ATC Labor
ATSC Hiring
25,000
# Candidate +
Developmental + Certified
ATCS on Payroll
Shortfall in
Certified ATCS
# ATSC
20,000
15,000
FAA-NATCA Target
Staffing Level
# Certified
ATCS
10,000
5,000
# Developmental
ATCS
#ATCS Attrition
= #ATCS Hired
ATSC in
Excess of
Target Level of
Staffing on
Payroll
0
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Year
16
Addressing the Challenges ?
• Are Challenges symptoms or causes?
– What are cause-effect relationships?
• What is the systemic structure of the industry?
– Can this knowledge explain behavior?
– Can this knowledge focus R&D?
– How do policies, regulations, and procedures affect
the system
– How does NGATS affect these success and
challenges
17
A member of
FAA Center of Excellence
2. Air Transportation System
CENTER FOR AIR TRANSPORTATION
SYSTEMS RESEARCH
Education
Analysis
Ed
ti
A l i and
d Research
R
h for
f the
th Next
N t
18
Air Transportation System
• Air Transportation System is …
– Layers of networks
– Networks composed of agents
– Agents:
• Distributed
• Autonomous
• Networks and agents operating with own objective
functions
– Reinforce/Undermine each other
• Networks exhibit:
– stochastic behavior
– operate in non-equilibrium state
• Economically
• System Performance
19
ATS Stakeholders
STAKEHOLDERS
Passengers and Cargo
•
•
Provide capacity for air transportation of
passengers and cargo
Scheduled Flights (routes, frequency
and aircraft)
Airports & Airways
•
•
Provide capacity for Airline’s Scheduled
Flights
Airways and their navigational aids,
Flightlevels, Runways, Gates, …etc.
Air Traffic Control
•
Provide sequencing and separation of
air traffic (flow)
Public Natural Resources
•
Optimize costs, time and reliability
Purchase air transportation services
Airlines
•
OBJECTIVE FUNCTIONS
Provides “natural resources” consumed
by air transportation
Profit
Marketshare in competitive marketplace
Maximize economies of scope and scale
Regional Economy
Effective Capacity
Congestion
Throughput (Delays)
Airports & Airspace Utilization
Accidents/Incidents
Workload
Capacity
Rate of Utilization
Rate of Replenishment
20
Relationship between Stakeholders
PUBLIC NATURAL RESOURCES
NOISE,
EMMISIONS
FROM TRAFFIC
AIR TRAFFIC CONTROL
SEQ. & SEP.
OF FLOW
OF TRAFFIC
CAPACITY FOR
ABSORBING NOISE
EMISSIONS FROM
TRAFFIC
FLOW OF
TRAFFIC
AIRSPACE & AIRPORTS
GATES,
RUNWAYS,
AIRWAYS,…
SCHEDULED
OPERATIONS
AIRLINES & OTHER USERS OF THE NAS
AVAILABLE
SEATS
UNCONSTRAINED
TRIP DEMAND
PASSENGERS & CARGO
AIR TRANSPORTATION SYSTEM
21
Variability in Demand
• Environment
• Demographics
S
Traffic Flow Service Providers:
Air Traffic Control (TFM, AOCs,
ARTCCs, TRACONs, Towers)
D
• Congestion
• Accident prevention
• Weather
• Budgets, Labor, Contract
• Technology
S
Air Transportation Service
Providers:
Airlines
D
• Seasonal, Day-of-Week, Hour-of-Day,
• Income Elasticity
• Price Elasticity
D
• Competition, Prices of competing modes
• Supply Chain price (e.g. Fuel)
• Technology & Labor costs
• Supply of resources (e.g. gates, aircraft)
S
Infrastructure Service Providers:
Airports and Airspace
Air Transportation Consumers:
Passengers & Cargo
AIR TRANSPORTATION SYSTEM
22
A member of
FAA Center of Excellence
3. ATS Dynamics
3-1) Time Constants
3-2) Demand and Capacity Balancing
Feedback Loops
3-3) System Equilibrium
CENTER FOR AIR TRANSPORTATION
SYSTEMS RESEARCH
Education
Analysis
Ed
ti
A l i and
d Research
R
h for
f the
th Next
N t
23
Dynamical System Model
Market Forces
Market Forces
Value of Substitute
Service/Product
Supply of
Service/Product
+
-
Control Law:
Price Setting
through “Price
Curve”
+
-
Plant:
Consumer Demand
through “Demand
Curve”
Demand
for
Service
Product
Phase-plane
Plant exhibits nonlinear dynamics
(e.g. price
elasticity)
Equilibrium
No-equilibrium
24
Time Constants ?
Changes in Pax & Cargo
demand
•
Changes in Airlines supply
– Routes, Frequency, Gauge
• τ = Weeks to 3 -6 Months
D
– τ = Weeks to Months
S
•
Traffic Flow Service Providers:
Air Traffic Control (TFM, AOCs,
ARTCCs, TRACONs, Towers)
:
e Providers
ic
rv
e
S
re
Infrastructu
d Airspace
Airports an
Changes in A&A supply
– Sector changes
• τ = 3-6 months, 2 years
• τ = 10 – 30 years
•
Changes in ATC
– Staffing, Sectors
D
– Runways, gates, routes, Crz
FLs
ice
tion Serv
ta
r
o
p
s
n
a
Air Tr
s:
Provider
Airlines
S
•
S
• τ = 3 years
D
– Increased Fleet
Air Transportation Consumers:
Passengers & Cargo
• τ = 7 – 10 years
AIR TRANSPORTATION SYSTEM
25
Pax v. Seats Feedback Loop (1)
•
Strong feedback-loop
– Airfares set by airlines based on
demand for available seats
– Scarce resources (seats) result
in higher prices
Air Transportation Service Providers:
Airlines
Seat Supply
P
D
•
Feedback distorted by:
– Hyper-competition for
marketshare
– Monopolies on given routes
– Bankruptcy protection
– Network integrity
– Use-it-or-Lose it rules
•
Innovation Rate
– Very high
Airfare
Enplanements
– Loop has no external costs
Seats
• Revenue Management
D
P
Air Transportation Consumers:
Passengers & Cargo
AIR TRANSPORTATION SYSTEM
26
Flights v. Slots - Feedback Loop (2)
•
Weak feedback-loop
– Fees and Taxes based on:
• Revenue-neutrality
• Congestion (delay) costs
Air Transportation Infrastructure:
Airports & Airspace
Airspace &
Airports
Supply
P
– Weak
Airspace &
Airports
Capacity
D
• Fees &
Taxes
• Congestion
Costs to
airlines
• not responsive to market
•
Feedback distorted by:
– Absence of value of scarce
resource
•
Scheduled
Flights
– FAA F&E Budget
– Costs borne by (third party)
passengers
– Administrative measures
(slots, uni-laterals)
Innovation Rate
– Slow
– Negligible productivity
improvements at chokepoints
D
P
Air Transportation Service Providers:
Airlines
27
Slots v. Sep. - Feedback Loop (3)
– Contract, budget & staffing
plans determine staffing
levels based on forecasts
– Costs borne by (third party)
passengers (& taxpayers)
– Absence of value of scarce
resource
P
D
• Salaries,
budgets,
staffing
plans
• Contract
• Innovation Rate
– Slow
– Negligible productivity
improvements
• # aircraft per sector in 15
mins
Demand for
Flow
Control
• Feedback distorted by:
Flow Supply
Flow
Control
Productivit
y
• Weak feedback-loop
Traffic Flow Service Providers:
Air Traffic Control (TFM, AOCs,
ARTCCs, TRACONs, Towers)
D
P
Air Transportation Infrastructure:
Airports & Airspace
28
Slots v. NPR - Feedback Loop (4)
• Feedback distorted by:
– Absence of value of scarce
resource
– Global and local issue
PNR Supply
P
D
• Costs borne
by public
Demand for
Natural
Resources
– Capacity and Rate-ofReplenishment is under debate
– Costs borne by (third party)
public
– Regulations (noise abatement,
engine upgrades)
Public Natural Resources:
Air & Water Quality, Noise
Supply of
Natural
Resources
• Weak feedback-loop
D
• Innovation Rate
– High in some areas
– Driven by forces other than
scarcity of resources
P
Air Transportation Infrastructure:
Airports & Airspace
29
Feedback Loops
Infrastructure:
Airports and
Airspace
Balancing Mechanism
(staffing, salaries, bu
dgets)
Demand
Service
Providers:
Traffic
Flow
Natural
Resource
Providers:
Public/Gov
ernment
Supply
Service
Providers:
Airlines (Flights)
ism
Balancing Mechan
tax,
t
ke
tic
,
es
fe
(landing
admin measures)
Infrastructure:
Airports and
Airspace
Supply
Demand
Consumers:
Passengers &
Cargo (Trips)
Demand
Balancing Mechan
ism
(airfares, subsidies
)
Service
Providers:
Airlines (Seats
)
Supply
Excess
Demand
Excess
Supply
AIR TRANSPORTATION SYSTEM
30
Feedback Loops
•
Absence of feedback loops prevents:
1. Maximizing use of scarce resources
•
Exacts costs on third parties not part of loop (Golaszewski,
2005)
–
–
delays, workload, overtime, sick-leave, pollution
not a problem until resource becomes scarce
2. Signaling for improvement
•
•
Crisis results in action
75% of capacity rule (Miller and Clarke, 2005)
3. Innovation
•
•
•
absence of value of resource (given demand) prevents
return-on-investment (ROI)
absence of ROI prevents venture capital (VC)
absence of VC prevents best-minds from investing time &
energy
31
Equilibrium ?
• Contributes to challenges
to ATS
ice
tion Serv
ta
r
o
p
s
n
a
Air Tr
s:
Provider
Airlines
D
– Time constants
– Weak feedback loops
D
• ATS is unable to reach
and maintain equilibrium
due to:
S
– ability to anticipate
changes
:
e Providers
ic
rv
e
S
re
Infrastructu
d Airspace
Airports an
S
• No shortage/excess
D
– supply = demand
Traffic Flow Service Providers:
Air Traffic Control (TFM, AOCs,
ARTCCs, TRACONs, Towers)
S
• Equilibrium:
Equilibrium
Air Transportation Consumers:
Passengers & Cargo
AIR TRANSPORTATION SYSTEM
32
A member of
FAA Center of Excellence
4. Opportunities
CENTER FOR AIR TRANSPORTATION
SYSTEMS RESEARCH
Education
Analysis
Ed
ti
A l i and
d Research
R
h for
f the
th Next
N t
33
Opportunities
1. Create awareness and educate decisionmakers, media, and public
• Gross mis-understandings exist, must be
addressed
• NAS Strategy Simulator
2. Plans & budgets should incorporate
dynamics of model
• time-constants, equilibrium points
• Key metrics: (1) effective capacity, (2) demand
• 75% of capacity rule
34
Opportunities
3) Sustainability in future ATS (e.g. NGATS) can
be achieved by including in the specification of
mechanisms to
•
balance existing supply with demand
•
•
•
Establish property rights and liability (Coase, 1988)
signal need for capacity enhancement
incentivize innovation, renewal, & expansion
•
Airports & Airspace, Air Traffic Control, Public Natural
Resources
35
A member of
FAA Center of Excellence
Thank you
Questions?
Lance Sherry - lsherry@gmu.edu, 703-993-1711
CENTER FOR AIR TRANSPORTATION
SYSTEMS RESEARCH
Education
Analysis
Ed
ti
A l i and
d Research
R
h for
f the
th Next
N t
36