Jupiter TFM Simulation Environment

1
Jupiter TFM Simulation Environment
2
Topics Left to Impress You With
100
Number of Topics Left
0

• 0900

1640
3
What is Jupiter TFM Simulation Environment (JSE)?

• JSE is a proven, distributed, human-in-the-loop
  simulation environment that consists of two
  parts
• A core flight simulator (potential to simulate
  the entire NAS for a day or multiple days)?
• An emulator of the traffic flow management
  infrastructure
• Objectives of using JSE
• Evaluation and development of new concepts in
  congestion management
• Procedural change evaluation and validation
• Technology integration and evaluation without
  effecting the operational environment
• Engage key stakeholders through interactive war
  games
• Fast-time simulations for post analysis and
  reporting

4
JSE/ISE Operating Environment
5
JSEs Flight Simulation

• Starts with an historical snapshot of actual
  flight data
• We hold over 7 years of archived flight data
• Simulates the major phases/states of a flight
• Scheduled
• Filed
• Taxi Out/In
• Runway departure/arrival
• Ascent/Descent
• Cruise
• Uses queuing models at constrained NAS resources
• Runways (also currently developing a higher
  granularity simulation for the airport surface
  gates, taxiways, etc.)?
• Arrival fixes (FCA, FEA, and sector demand
  managed through human interaction)?
• MIT restriction impacts occur through integration
  with FACET
• Adds unpredictability by random variance in the
  gate departures

6
JSEs TFM Infrastructure Emulation

• Allows various users of the system to interact
  while the simulation is running
• Allows traffic managers, airline dispatchers, and
  researchers the ability to interact with the
  simulation
• Network connections exist for remote users to
  participate in HITLs
• Many of the standard TFM and CDM tools can
  connect to the simulation environment as they
  would to the operational environment
• Flight Schedule Monitor
• Enhanced Substitution Module
• Route Management Tool
• Real-time Flight Schedule Analyzer
• Departure Flow Management

7
JSEs TFM Infrastructure Emulation (Contd)?

• Runs many of the same algorithms that exist in
  TFMS today
• Contains Adaptive Compression logic
• CDM Messaging
• Substitution processing
• Pop-up flight EDCT assignments
• Uses an internal trajectory modeler to predict
  demand at points in the airspace, such as
  Centers, Sectors, and user-defined Flow
  Constrained/Evaluation Areas
• Compatible with TFM Modernization
• SEVEN
• Enhanced pop-up processing

8
System Architecture
9
JSE Tools and Technologies
RMT-R
FSM
10
User Interaction with JSE

• Traffic Management Initiatives
• GDP, GS, AFP
• Broadcasts out EDCTs
• Flow Constrained/Evaluation Areas
• Provides back-predicted entry times into airspace
  using its internal trajectory model
• Initial Flight Plans and Amended Flight Plans
• Introduce new flights using a flight plan
• Simulates flights on their revised trajectories
• CDM Messaging
• Updated gate times
• Substitutions (simplified and slot credit)?
• Flight cancellations and new flight creation

11
Successful Uses of JSE (2002-2008)?

• Slot Credit Substitutions
• From concept to operational deployment in 14
  months
• Vision 100 Collaborative Planning exercises
• General Aviation Airport Programs (GAAP)?
• Airspace Flow Programs
• Tremendous involvement from industry stakeholders
  (Bi-weekly HitLs during concept development)?
• Heavy emphasis on procedural development
• TFM Training
• Multiple scenarios
• Departure Flow Management
• User group HitLs
• SEVEN
• Ongoing user group HitLs leading to an
  operational deployment in 2010/2011

12
SEVEN System Enhancements for Versatile
Electronic Negotiation
13
Basic Idea

• Users submit flight plans with multiple routing
  options and update these options as often as
  needed
• FAA identifies areas of interest/concern by
  creating interactive dynamic flight lists
• As conditions change, the demand allowed through
  the constrained area is adjusted up or down by
  checking/unchecking flights on list
• Unchecked flights are placed on highest available
  priority option that drop them from the list or
  if no such option exists then a flight is delayed
  on ground
• System impact assessment capabilities allow the
  FAA to model impacts on the NAS of different
  actions
• Manages uncertainty by being extremely agile
• Allows the easy recovering of resources if
  conditions improve/change reducing need for Wait
  and See
• Gets the FAA out of the business of finding
  routes (new paradigm)?

14
Multiple Routing Options

• Prioritized routing options consist of different
  combinations of routes, altitudes, speeds, and
  ptimes
• Users update/adjust list of options dynamically
  as needed
• Tools like ROG could help users to develop a set
  of feasible options that avoid constraints for a
  particular flight
• Default last choice option wait on the ground
• If ground delay issued then user can still update
  list of options to avoid constraint and remove
  ground delay
• Options priorities remembered until updated by
  user (or expire)?
• Only user submitted options (or ground delay)
  would be considered, thus avoiding need for
  electronic exception

• Example option set
• Fly route XYZ (through FEA) as planned
• Same route, but leave 20 minutes late (Minimize
  fuel burn)?
• Same route, but lower cruise altitude to fly
  under FEA and leave on time
• Fly route ABC (5 longer route (a CDR) around
  FEA), at altitude H (optimum?), at slightly
  slower speed (fuel efficient), and leave on time

15
Interactive Dynamic Flight Lists

• Lists can be created for any NAS resource of
  interest
• Lists are dynamic update as changes occur
• Lists are interactive demand is managed from
  within the list
• Flights that are checked can use the resource
• Unchecked flights are placed on highest priority
  option that avoids the resource. If no such
  option exists, then a flight is delayed on ground
  (if not yet active)?

• Auto-suggest algorithms could automate which
  flights are removed/added to list based on
  rationing/equity concerns (dial-up/dial down
  concept)?
• Lists allow for the re-capture of capacity when
  conditions unexpectedly improve/change (dial-up
  concept)?

16
Collaborative Routing Resource Allocation Tool
(CRRAT)?

• Functionality
• FACET interaction
• and HITL usage

17
CRRAT Overview
18
What is CRRAT?

• An algorithm to assign scarce (constrained) NAS
  resources to flights and/or carriers, given
  capacities on those constraints and requests to
  use them
• Both equity and efficiency of solution are
  considered
• Fast run time to this NP-hard scheduling problem
• Solving NAS-wide problems (several hundred
  resources and thousands of flights) in seconds to
  minutes.
• Highly flexible and configurable
• Resources could be airports, runways, taxi
  spots, waypoints, sectors, etc. It doesnt
  know/care what the resources are
• Collaborative designed to dovetail with user
  needs and input
• Multiple route options per flight

19
Multi-Objective Purpose

• Allocate scarce NAS resources
• Maximize (i) system efficiency, (ii) airline
  efficiency, and (iii) equity
• subject to constraints that
• Each flight be assigned a viable path and
  departure time
• For each time period, and for each resource, the
  capacity is not violated

20
Sample Applications

• NAS-wide flow control
• E.g. All sectors and airports
• Flow Constrained Area (FCA)
• Sectors not recommended
• GDP airports (trivial application)?
• Large-scale weather fronts
• e.g. that restrict flow to the east coast of the
  United States, storm systems that block passage
  across Cleveland Center, and the TFM desire to
  reduce arrival flow to the New York City area.
• Multi-fix GDP problem
• Each fix is established as a resource, as well as
  the airport itself.
• Each flight intending to arrive at the GDP
  airport requires use of exactly one of the
  arrival fix resources and the airport resource.

21
Configurable Priority Rules

• Examples of flight prioritization rules that can
  be achieved
• Grover-Jack (First-come First-served)?
• Ration-by-Schedule (RBS)/(First-scheduled
  First-served)?
• Accrued Delay
• Time-ordered Accrued Delay (TOAD)?
• Random Flight Selection
• And many more
• Best algorithms achieving 15-30 less total delay
  than RBS

By rearranging the order of the priority modules,
variations on the algorithm can be created by the
user. (Mr. Potato Head, if you will)?
22
CRRAT Software Summary
23
Input / Output
Input
Algorithm
Output
Demand forecast Proposed routes and depart times
for flights.
CRRAT Dispatching Rule Process flights one at
a time (resource hopping). Apply flight
prioritization rules Set control actions on each
flight.

• Controls
• Controlled departure time
• Controlled route

Metrics Equity Total delay Etc.
Capacity forecast Max arrival rate for airports
and airspace sectors in each future t.
Repeat
24
CRRAT Library

• Generic Java library with the core resource
  allocation capabilities from CRRAT
• Multi-resource, interdependent scheduling
• Multiple routes per flight
• Stable, reusable software, currently used by
• SEVEN
• Command and Control SBIR
• CRRAT

25
FACET Interaction

• CRRAT exists as an application with NASAs Future
  ATM Concepts Evaluation Tool (FACET)?

CRRAT
FACET
26
CRRAT HITLs

• CRRAT played a large role in the launching of the
  Airspace Flow Program (AFP) Traffic Management
  Initiatives
• AFP supporting analysis in 2004-2005
• Demonstrated the failings of GDPs in support of
  SWAP and proffered AFPs as an alternative
• CRRAT served as a prototype AFP system,
  integrated with JSE
• Demonstrations to senior FAA decision makers in
  early 2005
• AFPs went operational in 2006

27
CRRAT HITLs
28
CRRAT HITLs
29
Backup Slides
30
Brief History

• Phase 1 and Phase 2 SBIR project sponsored by
  NASA (2002-2004)?
• Major features and methodology developed
• Elaboration of Jason Burke MS thesis (UMD)?
• Further embellishment through C2 SBIR and various
  FAA projects
• Used in support of initial AFP analysis and
  demonstrations (2005)?

31
More Detailed Input / Output
32
Core Algorithm
33
Output (Decision Variables)?

• Each flight f ? F receives a path (sequence of
  resources) that it will use along with the time
  it should be at each resource

34

• The objective is to create an algorithm that
  allocates en route resources in a strategic,
  CDM-compatible manner

35
Alternate Approaches We Considered
Academic (optimization models)?
Industry software and tools

• CDM-FCA Working Subgroup
• Mitre-CAASD CRCT offspring
• MetronAV/NASA CRRAT
• EUROCONTROL) Computer Aided Route Allocation
  Tool (CARAT) is an optimization rerouting tool

• Bertsimas-Stock (MIT)?
• Goodhart Yano (UCB)?
• Ray Staats (AFIT)?
• NEXTOR-UMD ERAP

• Our work captures the desirable properties of a
  rationing algorithm

36
Incorporating Routing Preferences

• Three routes available to a flight. Air delay
  (e.g. extra miles flown over the shortest middle
  route) can be computed in advance.
• Once potential ground delays are known for each
  route, then cost of each route for the carrier
  can be computed.

Upper route is cheapest

• All computations but the ground delays can be
  computed and supplied to an automation in
  advance.
• This way, an FAA algorithm can choose the most
  appropriate route for each flight, taking carrier
  preferences into account.