Upload
tanith
View
29
Download
0
Embed Size (px)
DESCRIPTION
DARPA. CoAX Stand-alone Contributions DARPA Briefing - November 2000 Dartmouth College, UMichigan, MIT Sloan, Coalition Agents eXperiment (CoAX) http://www.aiai.ed.ac.uk/project/coax/. Stand-alone Contributions. Dartmouth Field Observation Agent MIT Robustness Service - PowerPoint PPT Presentation
Citation preview
CoAX Stand-alone ContributionsCoAX Stand-alone ContributionsDARPA Briefing - November 2000DARPA Briefing - November 2000
Dartmouth College, UMichigan, MIT Sloan, Dartmouth College, UMichigan, MIT Sloan,
Coalition Agents eXperiment (CoAX)Coalition Agents eXperiment (CoAX)http://www.aiai.ed.ac.uk/project/coax/http://www.aiai.ed.ac.uk/project/coax/
DARPADARPA
CoAX /Tech Briefing - 2
Stand-alone Stand-alone ContributionsContributions
Dartmouth Field Observation Agent Dartmouth Field Observation Agent MIT Robustness ServiceMIT Robustness Service Michigan Coordination Planning AidMichigan Coordination Planning Aid
CoAX /Tech Briefing - 3
• ActComm Project
• Dartmouth, Harvard, RPI, Illinois, ALPHATECH, Lockheed Martin
• Department of Defense Multidisciplinary University Research Initiative
• Developing a system to provide network access to soldiers in the field
• CoAX Goal
• Demonstrate the ease with which the large ActComm “legacy” system can be integrated with the rest of CoAX via the DARPA CoABS Grid
Field Observations (Dartmouth)
CoAX /Tech Briefing - 4
Field Observations (Dartmouth)
• Team of soldiers
• PDA’s
• Ad-hoc wireless networking
• Soldiers make observations.
• Ground and air traffic
• Personnel and equipment
• Buildings and other structures
• Observations fed into battle-planning systems (e.g., MBP) through the CoABS Grid.
• In the demo, a team of CoAX soldiers will make observations to correct Gao misinformation.
CoAX /Tech Briefing - 5
Observations
ObservationAgent
D’Agents API
GridAPI
I see a tank!
ObservationViewer
MBP
(9-month demo - standalone)
(18-month demo - integrated)
Query/Response
Registration/Update Stream
Field Observations (Dartmouth)
CoAX /Tech Briefing - 6
Field Observations (Dartmouth)
29-SEP-2012 13:47.56 OBSERVATION 0018 VEHICLE Observer : 16.35 N, 35.28 E, Elevation 530 m Sightline: 270 deg, 0 deg down, 2000 m Vehicle : Gao, flatbed truck, 3 axles, heading: 180, speed: 60 km/h Note : 12 soldiers in flatbed
CoAX /Tech Briefing - 7
The Challenge: The Challenge: Robust Agent Robust Agent CoalitionsCoalitions
Coalitions are open systemsCoalitions are open systems Dynamic membership, often novel partnersDynamic membership, often novel partners
Agents in open systems will be unreliableAgents in open systems will be unreliable Intermittent bugs (3 per 1000 lines in the best Intermittent bugs (3 per 1000 lines in the best
crafted code) as well as the possibility of malicecrafted code) as well as the possibility of malice Infrastructures can be unreliableInfrastructures can be unreliable
Current failure tolerance approaches are Current failure tolerance approaches are insufficientinsufficient Assume closed systems (e.g. mirroring)Assume closed systems (e.g. mirroring) Full rollbacks are unnecessarily inefficient for agentsFull rollbacks are unnecessarily inefficient for agents
CoAX /Tech Briefing - 8
The MIT Robustness The MIT Robustness ServiceService
Monitors agent ‘health’ via pollingMonitors agent ‘health’ via polling Responds to agent failure via intelligent task Responds to agent failure via intelligent task
cancellation & task re-announcementcancellation & task re-announcement Maintains reliability information (for failure Maintains reliability information (for failure
avoidance)avoidance) Designed for open systems - makes minimal Designed for open systems - makes minimal
assumptions about agentsassumptions about agents
CoAX /Tech Briefing - 9
A Working Grid A Working Grid ServiceService
MessageLog
RobustnessService
EH API
Transparently infers Transparently infers commitment structurescommitment structures
Assumes (some) agents Assumes (some) agents support (some of ) EH support (some of ) EH APIAPI Polling (backup: Polling (backup:
existing Grid is-alive? existing Grid is-alive? method)method)
Task re-announceTask re-announce Cancel-taskCancel-task
CoAX /Tech Briefing - 10
CoAX /Tech Briefing - 11
Benefits Validated Benefits Validated EmpiricallyEmpirically
Up to 3x speedup and 8x reduced variability Up to 3x speedup and 8x reduced variability vs. standard timeout-retry approachvs. standard timeout-retry approach
Benefits increase with task complexity Benefits increase with task complexity (decomposition tree height) and with level of (decomposition tree height) and with level of EH API supportEH API support
http://ccs.mit.edu/klein/papers/ASES-WP-2000-http://ccs.mit.edu/klein/papers/ASES-WP-2000-05.ps05.ps
CoAX /Tech Briefing - 12
Michigan Multilevel Michigan Multilevel Coordinator AgentCoordinator Agent
Analyses the alternative plan spaces of coalition functional Analyses the alternative plan spaces of coalition functional teams that plan independently and act asynchronouslyteams that plan independently and act asynchronously
Works top-down with plans chosen by teams to predict Works top-down with plans chosen by teams to predict unintended interactions (resource contentions; friendly fire).unintended interactions (resource contentions; friendly fire).
Identifies candidate resolutions (timing or action Identifies candidate resolutions (timing or action constraints).constraints).
Notifies process panel of possible plan conflicts and Notifies process panel of possible plan conflicts and computed workarounds.computed workarounds.
Operationalizes/enforces coordination decisions selected.Operationalizes/enforces coordination decisions selected. Given more time, isolates and resolves conflicts more Given more time, isolates and resolves conflicts more
precisely and efficiently.precisely and efficiently. Allows planning and coordination decisions to be postponed Allows planning and coordination decisions to be postponed
until runtime conditions become better known.until runtime conditions become better known. Packaged as a Grid-aware component that will be Packaged as a Grid-aware component that will be
proactively executing and will be utilized by the AIAI proactively executing and will be utilized by the AIAI Process Panel.Process Panel.
CoAX /Tech Briefing - 13
Potential plan conflicts include friendly fire in TEZ on ArmyDiv2, destruction of roads through E that ArmyDiv2 might need, and contention for sea and rail transport among army divisions and logistics.
Michigan Coalition Michigan Coalition Coordination ExampleCoordination Example
Forces begin at aircraft carrier ACAirforce sorties to C, E, & Q for Total Exclusion Zone (TEZ)Logistics delivers humanitarian aid to refugees at F and RArmyDiv1 occupies X to prevent Agadez forces from reaching and inciting refugees at RArmyDiv2 crosses TEZ to occupy Y to monitor for Gao crossings
CoAX /Tech Briefing - 14
Coordinated PlansCoordinated PlansHierarchical plan coordination incrementally Hierarchical plan coordination incrementally
recommends coordinated plans that are recommends coordinated plans that are increasingly detailed and parallelizedincreasingly detailed and parallelized
Fly sorties
Move to X
Time = 6500.02 cpu
sec.
Time = 5000.38 cpu sec.
Time = 4256.04 cpu sec.
Move to Y
LogisticsAirforce
Army Div 1Army Div 2
Move C1 R
Fly sortiesACP
Move to Y
LogisticsAirforce
Army Div 1Army Div 2
P P
P Z Z X
Move C1 R, C2 F
Fly sortiesACP
Move to Y
LogisticsAirforce
Army Div 1Army Div 2
P P
P Z Z X
Move C2 F
P P Move C1 R, C2 F
CoAX /Tech Briefing - 15
Michigan Multilevel Michigan Multilevel Coordinator AgentCoordinator Agent