Airport Operations Analysis
Using Fast-Time Simulation Models
Dr. Alexander Klein
Jianfeng Wang
Stephen Szurgyi
CENTER FOR AIR TRANSPORTATION
SYSTEMS RESEARCH
GMU Airport Operations Research
Capacity
Safety
Economic Analysis
Wide range of models & tools
Stochastic Simulation
•
Identifying best solution in a fastest way
99% Confidence Intervals
Bag
Throughput
Bag
Throughput
#1 #2 #3 #4 #5
Equipment Solutions
Single run
2
#1 #2 #3 #4 #5
Equipment Solutions
Multiple runs
CATSR
LaGuardia Airport – “Upgauging”
CATSR
Problem
•
One of nation’s most congested airports
•
Slot controlled
•
Large proportion of regional jets (RJs)
•
FAA and NYPA looking for ways to increase pax throughput
from 25 to 30M / year without increasing congestion
Proposed action
•
Incentivize air carriers to “upgauge” to larger aircraft
Question posed to research community
•
Would upgauging negatively affect LGA operations/delays?
Must look at Runway, Taxiway and Gate capacity
3
Simulation Scenarios
RWY Config
CATSR
Day
061404 (ETMS)
A22/D13
041905 (GRA)
081105 (ETMS)
061404 (ETMS)
A22/D31
041905 (GRA)
081105 (ETMS)
4
Fleet mix
Baseline
12% less RJs
25% less RJs
Baseline
Upgauged (50% less RJs)
Upg (757s to A321s)
Baseline
12% less RJs
25% less RJs
Baseline
12% less RJs
25% less RJs
Baseline
Upgauged (50% less RJs)
Upg (757s to A321s)
Baseline
12% less RJs
25% less RJs
CATSR
5
CATSR
6
Summary of Results
Run
Config
Number
of
Schedule Flights
in
TAAM
Fleet Mix
CATSR
Total
Avg
Avg
Est'd
Available Seats Annual Pax Delay
per
per
Seats
excluding
GA
Flight
per Day Flight
Avg Average
Delay Standoff
per Time per
Seat
Flight
Est'd
Carriers Most
PeakAffected by
Time
Gate Shortages
Gate
Shortage
Minutes
1
2
3
4
5
6
ETMS
6/14/2004
(filled to
75 ops/hr)
A22/D13
(Peak
sustained GRA 0419
AAR/ADR
= 39.5)
7
8
ETMS
8/11/2005
9
10
11
12
ETMS
6/14/2004
(filled to
75 ops/hr)
1206
Baseline
111279
92
24,795,807
16.0
15.4
0.9
1
None
1208
12% RJ to NB
114456
95
25,461,499
17.8
17.6
0.9
1
None
1208
25% RJ to NB
118192
98
26,292,597
16.9
16.8
0.7
1
None
1178
Baseline
113142
96
25,810,172
11.4
10.5
1.0
1
None
1190
Upgauged
135165
114
30,523,169
15.2
14.9
1.5
2
COM, EGF, USA
Upg, 757to321 133075
112
30,051,202
13.0
12.9
1.6
2
COM, EGF, USA
1190
1222
Baseline
113851
93
25,036,751
19.0
17.8
2.6
4
USExp, COM
1222
12% RJ to NB
117305
96
25,796,314
17.5
16.6
2.7
4
USExp,EGF,COM
1222
25% RJ to NB
122567
100
26,953,470
18.6
18.0
2.6
4
USExp,EGF,COM
1206
Baseline
111279
92
24,795,807
23.5
22.4
1.7
2
COM, EGF
1208
12% RJ to NB
114456
95
25,461,499
25.1
23.8
1.5
2
COM, EGF
1208
25% RJ to NB
118192
98
26,292,597
27.0
26.1
2.2
3
COM, EGF
1178
Baseline
113142
96
25,810,172
16.0
14.2
0.9
1
None
1190
Upgauged
135165
114
30,523,169
23.0
21.8
2.5
4
COM, EGF, USA
Upg, 757to321 133075
112
30,051,202
21.2
20.0
2.5
3
COM, USA, EGF
A22/D31
13
(Peak
14 sustained GRA 0419
AAR/ADR
15
= 38)
1190
16
1222
Baseline
113851
93
25,036,751
27.8
25.1
3.0
4
USExp,EGF,COM
1222
12% RJ to NB
117305
96
25,796,314
27.3
25.0
2.9
4
USExp,EGF,COM
1222
25% RJ to NB
122567
100
26,953,470
24.8
22.7
2.4
3
USExp,EGF,COM
17 7
18
ETMS
8/11/2005
Gate Activity (TAAM Simulation) - excerpt
Example: A22/D13, 04/19 (GRA) Upgauged
6 am
gtCTBA02
gtCTBA03
gtCTBA04
gtCTBA05
gtCTBA06
gtCTBA07
gtCTBB01
gtCTBB02
gtCTBB03
gtCTBB04
gtCTBB05
gtCTBB06
gtCTBC01
gtCTBC03
gtCTBC04
gtCTBC05
gtCTBC08
gtCTBC09
gtCTBC10
gtCTBC11
gtCTBC12
gtCTBD01
gtCTBD02
gtCTBD03
gtCTBD04
gtCTBD05
gtCTBD06
gtCTBD07
gtCTBD08
8
gtCTBD09
. . . . . . . . . . . .
COA
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
ACA
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
MDW
. . . . . . . . . . . .
COA
. . . . . . . . . . . .
MDW
. . . . . . . . . . . x
ATA/FFT
. . . . . . . . . . . x
Spirit, JBU
. . . . . . . . . . . x
. . . . . . . . . . . .
AAL/EGF
. . . . . . . . . . . .
. . . . . Gate
. . . . . . .
. . . . . utilization
. . . . . in. .
.UAL
. . . . TAAM
. . . is. only
. . .
. . . . . an
. . . . . . .
. . . . . approximation
. . . . . . .
. . . . . of
. a
. real
. . day
. . .
. . . . . of
. operation.
. . . . . .
. . . . . . . . . . . .
. . . . . This
. . plot
. . is
. a. .
. . . . . simplified
. . . . . . .
. . . . . representation
. . . . . . .
.AAL/EGF
. . . . of
. continuous
. . . . . .
. . . . . gate
. . . . . . .
. . . . . occupancy.
. . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . x
.
.
.
x
.
.
.
x
x
x
x
x
.
.
x
x
.
x
.
.
x
x
x
.
.
.
.
.
x
x
10 am
.
x
.
x
.
.
x
x
x
x
x
x
x
.
x
x
.
x
.
.
x
x
x
x
x
x
.
.
x
x
.
x
x
x
x
.
x
x
x
x
x
x
x
x
x
x
.
.
.
.
x
x
x
x
x
x
x
x
x
x
.
x
x
x
x
.
.
x
x
x
x
x
x
x
x
x
.
.
.
.
x
x
x
x
x
x
x
x
x
x
.
x
.
x
.
x
.
x
x
x
x
x
x
x
x
x
.
x
.
.
x
x
x
x
x
x
x
.
x
x
.
x
x
x
.
x
x
.
x
x
.
x
x
x
x
.
.
x
.
.
x
x
x
x
x
x
x
.
x
x
.
x
x
x
x
x
x
x
x
x
.
x
x
x
x
.
.
.
.
.
x
.
x
x
x
x
x
.
x
.
.
x
.
x
x
.
x
x
x
x
.
x
x
x
.
x
x
.
.
.
x
x
x
x
x
x
x
.
x
x
.
x
.
.
x
.
.
x
x
x
x
x
.
x
.
x
x
x
.
.
x
x
x
x
.
x
x
.
x
x
2 pm
.
x
.
x
x
x
.
x
.
x
x
x
x
x
.
x
.
x
.
.
x
x
x
x
x
.
x
.
.
x
.
x
.
x
.
x
.
x
x
x
.
x
x
x
x
x
.
x
.
.
.
x
x
x
x
.
x
.
x
x
.
x
x
x
x
.
x
x
x
x
.
x
x
.
x
.
.
.
.
.
x
x
x
x
x
.
x
.
x
x
.
x
x
x
x
.
x
x
x
x
x
x
x
x
.
.
x
x
.
.
x
x
.
x
x
x
.
.
.
x
.
x
.
x
x
x
.
x
x
.
x
x
x
x
x
x
x
x
.
.
x
.
x
x
.
x
.
.
x
x
.
x
.
x
x
x
.
x
x
x
x
x
.
x
x
x
.
x
.
x
x
x
x
x
.
x
x
x
x
x
.
x
.
x
x
.
x
x
x
x
.
.
x
x
.
x
x
x
.
x
x
x
.
x
x
x
x
x
.
x
.
x
.
.
x
.
x
x
.
x
.
.
x
x
.
x
x
x
.
.
x
x
.
.
x
.
x
x
.
x
CATSR
6 pm
.
x
.
.
x
.
.
.
.
x
x
x
x
x
.
x
.
x
.
.
x
.
.
x
x
.
x
x
.
.
.
x
.
x
x
.
.
.
x
x
x
x
.
x
x
x
.
x
.
.
x
.
x
x
.
x
x
x
x
x
.
x
.
x
x
.
.
x
x
x
x
.
x
x
x
x
x
.
.
x
x
x
x
x
.
x
.
x
x
x
.
x
.
x
.
.
x
x
x
x
x
x
x
x
x
x
x
.
.
x
x
x
x
.
x
x
.
x
.
x
x
x
x
.
x
.
x
x
x
x
x
x
x
x
.
x
x
.
.
x
x
.
.
x
x
.
.
x
.
.
x
x
x
.
x
.
.
.
x
x
.
x
.
x
.
x
.
x
.
x
x
x
.
x
.
.
x
.
x
x
x
x
x
.
.
.
x
.
x
x
.
x
.
x
.
.
x
x
x
.
x
x
.
.
.
.
x
x
x
x
x
x
x
.
.
x
x
x
x
x
.
.
x
.
x
.
x
x
x
x
.
x
x
.
x
.
x
x
.
x
10 pm
.
x
.
.
.
x
.
x
.
x
x
.
x
x
x
.
.
x
x
x
.
x
x
x
x
x
x
.
.
x
.
x
.
.
x
x
.
x
x
x
x
x
.
x
.
x
.
x
x
x
x
x
x
x
x
x
x
x
.
x
.
x
.
x
x
x
x
x
x
x
x
x
x
x
x
x
.
.
.
x
x
x
x
.
x
.
x
x
x
x
.
x
x
x
.
x
x
x
x
x
x
x
x
x
x
x
x
x
.
x
x
.
x
x
x
.
x
.
x
x
.
x
x
.
x
.
x
x
.
x
x
x
x
.
x
x
x
x
.
.
x
x
x
x
x
x
.
.
x
x
.
x
x
.
x
.
x
x
x
x
.
x
x
x
x
.
.
x
.
.
x
x
x
x
x
x
.
.
x
x
.
x
x
.
.
.
x
x
x
x
.
x
x
x
x
.
.
x
.
.
x
x
x
x
.
x
.
.
.
x
.
.
.
.
x
.
.
.
x
x
x
x
.
x
x
x
.
.
.
.
x
x
x
x
.
.
.
.
.
x
.
.
.
.
x
.
.
x
.
x
x
x
.
.
x
x
.
.
.
.
.
.
x
.
.
.
.
.
.
x
.
.
.
.
.
.
.
x
.
x
.
x
.
.
x
.
.
.
.
.
.
.
.
.
.
.
.
.
.
x
.
.
.
.
.
.
.
.
.
.
.
x
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Average Delay & Standoff Time / Flight
04/19/05 Sample (GRA)
CATSR
A22/D31, 04/19 (GRA 2005)
A22/D13, 04/19 (GRA 2005)
25.0
2.0
10.0
1.0
5.0
0.0
0.0
Baseline
Upgauged
Upg, 757to321
4.0
20.0
Delay
15.0
9
25.0
Standoff Time
Delay
3.0
5.0
Avg Standoff Time / Flight
4.0
20.0
Avg Delay / Flight, mins
30.0
5.0
3.0
15.0
2.0
10.0
1.0
5.0
0.0
0.0
Baseline
Upgauged
Upg, 757to321
Standoff Time
Avg Delay / Flight, mins
Avg Standoff Time / Flight
30.0
Understanding Differences in Delays
Impact of Schedule Lumpiness (1 of 2)
50
50
ARR
DEP
45
40
35
35
30
30
25
25
Demand- Base
20
20
15
15
10
Demand- Upgauged
10
80000
80000
5
Standoff (secs)
Gate (secs)
5
70000
70000
40000
40000
30000
30000
20000
20000
10000
10000
0
00
0:
:0
22
4::0
00
23 0
5::0
000
06:
:000
0
7:
00
8:
00
9:
00
10
:0
0
11
:0
0
12
:0
0
13
:0
0
14
:0
0
15
:0
0
16
:0
0
17
:0
0
18
:0
0
19
:0
0
20
:0
0
21
:0
0
22
:0
0
23
:0
0
00
21
:
00
20
:
00
19
:
00
18
:
00
17
:
00
00
16
:
00
15
:
00
14
:
00
13
:
00
12
:
00
11
:
10
:
0
9:
0
0
8:
0
0
0
7:
0
6:
0
0
0
5:
0
0
4:
0
0
:0
50000
23
:0
21
50000
010
Taxiway (secs)
Sequencing (secs)
0 00 00
00 0(secs)
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 Runway
0
4: 5: 6: 7: 8: 9: 10: 11: 12: 13: 14: 15: 16: 17: 18: 19: 20: 21: 22: 23:
600000
0
0
:0
20
19
:0
0
0
0
:0
:0
18
17
:0
0
0
:0
16
15
14
:0
0
0
:0
13
12
:0
0
0
0
:0
:0
11
10
9:
00
00
8:
7:
00
00
00
6:
00
5:
60000
Delays- Upgauged
0
Delays- Base
22
0
4:
ARR
DEP
45
40
CATSR
Understanding Differences in Delays
Impact of Schedule Lumpiness (2 of 2)
Modified, lumpier
schedule, larger
fleet: delay increase
is noticeable
Same schedule, larger fleet: delay
increase is small
15.0
2.0
10.0
1.0
5.0
0.0
0.0
Baseline
12% RJ to NB 25% RJ to NB
25.0
4.0
3.0
15.0
2.0
10.0
1.0
5.0
0.0
0.0
Baseline
12% RJ to NB
25% RJ to NB
Avg Delay / Flight, mins
30.0
5.0
Avg Standoff Time / Flight
20.0
Delay
3.0
Standoff Tim e
Delay
20.0
11
25.0
4.0
5.0
4.0
20.0
3.0
15.0
2.0
10.0
1.0
5.0
0.0
0.0
Baseline
Upgauged
Upg, 757to321
Standoff Time
25.0
Avg Delay / Flight, mins
Avg Standoff Time / Flight
30.0
5.0
Delay
Avg Delay / Flight, mins
Avg Standoff Time / Flight
Standoff Time
30.0
A22/D31, 04/19 (GRA 2005)
A22/D31, 0614 (2004)
A22/D13, 0614 (2004)
CATSR
Stochastic Simulations
Departure time, aircraft performance randomized
At congested airports like LGA, small
variations in traffic demand or
throughput can lead to large
fluctuations in delays
LGA Movements
Example for 04/19 schedule,
upgauged, A22/D31 runway
configuration, is shown
12
LGA Delays
CATSR
Average Delay Per Flight (04/19, GRA)
20 Randomized TAAM Runs for the 6 Scenarios
CATSR
LGA Avg Delay Per Flight - Six Randomized Multi-Runs (20 Iterations Each)
Primary parameter randomized is departure time
35.0
Delay per flight, minutes
30.0
25.0
A22D13 Base
20.0
A22D13 Upgauged
A22D13 757to321
A22D31 Base
15.0
A22D31 Upgauged
A22D31 757to321
10.0
5.0
0.0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Run #
Comparison with actual data: average delay/flight at LGA was
19.5 min in Aug 2005
13
LaGuardia Upgauging - Discussion
CATSR
Upgauging does not lead to significant delay increase – if
schedule “lumpiness” can be mitigated
At peak-demand times, LGA can be short of about 3-4 gates
• Gate shortages are highly “local” (carrier and time specific)
• Some un-used gates are always present, even at peak times
Optimized scheduling is key to reducing delays at a given
demand level
• Fine-tuning through simulation
Combined simulation of aircraft and passenger/bag movement
at LGA is highly desirable
14
Fast-Time Simulation Enables Broader Analyses
E.g. TAAM as a Carrier, Platform for Other Tools & Models
CATSR
Noise
analysis
Airport & airspace design tools
TAAM simulation
Scenario
generation
for real-time
simulators
Cost /
benefit
analysis
Passenger
and baggage
flow
simulation
15
Conflict analysis
Capacity statistics
CENTER FOR AIR TRANSPORTATION
SYSTEMS RESEARCH
CENTER FOR AIR TRANSPORTATION
SYSTEMS RESEARCH
30
SS
DSS
20
10
0:15:59
Time Period
23:45 - 00:00
25
23:00 - 23:14
50
22:15 - 22:29
30
21:30 - 21:44
60
20:45 - 20:59
35
20:00 - 20:14
70
19:15 - 19:29
40
18:30 - 18:44
80
17:45 - 17:59
45
17:00 - 17:14
90
16:15 - 16:29
50
15:30 - 15:44
100
14:45 - 14:59
23:30 - 23:44
23:00 - 23:14
22:30 - 22:44
22:00 - 22:14
21:30 - 21:44
21:00 - 21:14
20:30 - 20:44
20:00 - 20:14
19:30 - 19:44
19:00 - 19:14
18:30 - 18:44
18:00 - 18:14
17:30 - 17:44
17:00 - 17:14
16:30 - 16:44
16:00 - 16:14
15:30 - 15:44
40
14:00 - 14:14
13:15 - 13:29
12:30 - 12:44
11:45 - 11:59
11:00 - 11:14
10:15 - 10:29
09:30 - 09:44
08:45 - 08:59
08:00 - 08:14
0:10:39
07:15 - 07:29
0:13:19
15:00 - 15:14
SM
06:30 - 06:44
14:30 - 14:44
Modified Avg open
05:45 - 05:59
0
Time Period
14:00 - 14:14
Average Passenger
Delay
Opening additional 6-8
lanes reduces wait time
during off-peak periods
13:30 - 13:44
13:00 - 13:14
12:30 - 12:44
12:00 - 12:14
11:30 - 11:44
11:00 - 11:14
10:30 - 10:44
10:00 - 10:14
09:30 - 09:44
09:00 - 09:14
08:30 - 08:44
08:00 - 08:14
07:30 - 07:44
07:00 - 07:14
06:30 - 06:44
06:00 - 06:14
05:30 - 05:44
05:00 - 05:14
Avg open
05:00 - 05:14
18
04:30 - 04:44
04:00 - 04:14
Utilization (%)
IAD: Delay/Utilization Tradeoff
CATSR
Security Lane Utilization Comparison Between Baseline and DSS with Modified Schedule
DSS-Modified
Average Passenger Delay
20
15
10
DSS-Modified
0:18:39
5
0
0:08:00
0:05:20
0:02:40
0:00:00
From Individual Models to Toolsets
CATSR
From NAS-wide effects to individual airport models
First-class simulation tools
Stochastic modeling: scientific approach to using fast-time
simulation models
Plug-ins and data analysis tools built at GMU
“Air traffic analysis” (capacity/demand/procedures/expansion),
safety analysis, and economic analysis go hand-in-hand
Use of sophisticated tools and models to educate the next
generation of model users and analysts
19