ARCS Presentation 2008

27
An Anytime Winner Determination Algorithm for Time-Extended Multi- Robot Task Allocation Dr. Fang Tang & Spondon Saha Intelligent Robotics Lab California State Polytechnic University Pomona, CA

description

Was used at the ARCS\'08 conference held at Orlando, FL.

Transcript of ARCS Presentation 2008

Page 1: ARCS Presentation 2008

An Anytime Winner Determination Algorithm

forTime-Extended Multi-Robot Task

Allocation

Dr. Fang Tang & Spondon SahaIntelligent Robotics Lab

California State Polytechnic University Pomona, CA

Page 2: ARCS Presentation 2008

2

Outline

Introduction ASyMTRe-D Combinatorial Auctions

Simultaneous Task Allocation (STA)Performance MetricsFuture WorkConclusionReferencesQ&A

Intelligent Robotics Lab, Cal Poly Pomona

Page 3: ARCS Presentation 2008

3

Introduction: ASyMTRe-D

Automated Synthesis of Multi-robot Task solutions through software Reconfiguration.

ASyMTRe determines: a coalition of robots. from a heterogeneous group of robots. to carry out a given multi-robot task.

Intelligent Robotics Lab, Cal Poly Pomona

Page 4: ARCS Presentation 2008

4

Introduction: ASyMTRe-D

Heterogeneous group of robots??? Robots with different functional capabilities.

Multi-robot task??? A task that cannot be carried out by a single robot. Requires a “strongly cooperative” solution.

Intelligent Robotics Lab, Cal Poly Pomona

Page 5: ARCS Presentation 2008

5

Introduction: ASyMTRe-D

ASyMTRe Centralized algorithm. Runs on a central server. Aware of the number of robots and their individual capabilities.

ASyMTRe-D Distributed ASyMTRe designed using the Contract Net Protocol. Runs locally on each robot. Uses a group negotiation process to determine coalitions.

Intelligent Robotics Lab, Cal Poly Pomona

Page 6: ARCS Presentation 2008

6

Introduction: ASyMTRe-D

R1

R4 R3

R2R5Task A

Idle Robots

Intelligent Robotics Lab, Cal Poly Pomona

Page 7: ARCS Presentation 2008

7

Introduction: ASyMTRe-D

ASyMTRe-D takes into account: Task-Specific cost. Robot-inherent cost.

Determines a suitable mapping between: Task. Robots.

Intelligent Robotics Lab, Cal Poly Pomona

Page 8: ARCS Presentation 2008

8

Introduction: ASyMTRe-D

Problem: Can handle only one task at a time. Each round results in idle robots.

Desired behavior: Need to reduce instances of idle robots. Make idle robots work on other tasks (if any).

Intelligent Robotics Lab, Cal Poly Pomona

Page 9: ARCS Presentation 2008

9

Introduction: Combinatorial Auctions

Sequential Auctions.

Parallel Auctions.

Combinatorial Auctions.

Intelligent Robotics Lab, Cal Poly Pomona

Page 10: ARCS Presentation 2008

10

Introduction: Combinatorial Auctions

Item 1 Item 4Item 3Item 2 Item 5

Auctioneer

[1,2] [1,3,5] [3,5] [1, 4]

???

???

,[2,5]

Intelligent Robotics Lab, Cal Poly Pomona

Page 11: ARCS Presentation 2008

11

Introduction: Combinatorial Auctions

Complexity: A NP-complete problem. Similar to the Partitioning problem → NPC.

Dynamic Programming: Long execution time. Explores the entire solution space. Scales to only a small number of bids.

Intelligent Robotics Lab, Cal Poly Pomona

Page 12: ARCS Presentation 2008

12

Introduction: Combinatorial Auctions

Solution: An Anytime tree algorithm. Explores only the relevant solution space. Polynomial in the number of bids submitted.

Intelligent Robotics Lab, Cal Poly Pomona

Page 13: ARCS Presentation 2008

13

Introduction: Combinatorial Auctions

Bids:123451,21,3,51,42,53,5

Root

1,2

3,5 3

4 4

5

1,3,5

4

2

1,4

2,5 2

3 3,5 3

5

1

2,5 2

3 3,5 3

44 4

5

Intelligent Robotics Lab, Cal Poly Pomona

Page 14: ARCS Presentation 2008

14

Quick Review

ASyMTRe-D Drawbacks: Accepts tasks sequentially. Order of task execution determine total execution time. Idle robots are waste of resources.

Need: A task scheduler. Least robot idle-time → maximum utilization of resources.

Inspiration for STA: Combinatorial Auctions.

Intelligent Robotics Lab, Cal Poly Pomona

Page 15: ARCS Presentation 2008

15

Simultaneous Task Allocation (STA)

Task1 Task 2 Task 3

{R1,R2, $3}

R2

R3R1

{R2,R3, $4}

{R3, $5}

{R1,R2, $4}

{R2, $1}

{R3, $3}

{R1,R3, $4}

{R2,R3, $4}

Intelligent Robotics Lab, Cal Poly Pomona

Page 16: ARCS Presentation 2008

Simultaneous Task Allocation (STA)

`

Task Bids

1

{R1, R2, $3}

{R2, R3, $4}

{R3, $5}

2

{R1, R2, $4}

{R2, $1}

{R3, $3}

3

{R1, R3, $4}

{R2, R3, $4}

Root

R1, R2

R2, R3

R3

3 4 5 0

R3

6 3 4R1, R2

9

R2

6 5R1, R2

4R2

1R3

3 0

R2, R3

4R1, R3

4R1, R3

5

16

Intelligent Robotics Lab, Cal Poly Pomona

Page 17: ARCS Presentation 2008

17

Simultaneous Task Allocation (STA)

Will always return a partial answer if terminated early.

Each path is a set of disjoint coalitions → Partition.

All possible partitions of robots are considered.

Run time complexity depends on: Number of bids submitted. Number of tasks considered.

Intelligent Robotics Lab, Cal Poly Pomona

Page 18: ARCS Presentation 2008

18

Simultaneous Task Allocation (STA)

Task B

Task C

……

Task N

STA Algorithm

Task D

Task E Task F

Task A

Intelligent Robotics Lab, Cal Poly Pomona

ASyMTRe-D equipped…

Page 19: ARCS Presentation 2008

19

Performance (STA)

Test Environment: Dell Inspiron 6400. 1.73 GHz Pentium Dual Core Processors. 2 GB of RAM. Python.

Objective: To measure the execution time of the algorithm For increasing bids. For increasing task sizes.

Intelligent Robotics Lab, Cal Poly Pomona

Page 20: ARCS Presentation 2008

20

Performance (STA)

Random Distribution: Bids accepted → 10 - 100 Task Sizes → 5, 10, 15, 20 Robots → 10 Maximum allowed coalition size → 9

Uniform Distribution: Bids accepted → 10 - 50 Task Sizes → 5, 10, 15, 20 Robots → 10 Maximum allowed coalition size → 3

Intelligent Robotics Lab, Cal Poly Pomona

Page 21: ARCS Presentation 2008

21

Performance (STA)

Random Distribution

Intelligent Robotics Lab, Cal Poly Pomona

Page 22: ARCS Presentation 2008

22

Performance (STA)

Uniform Distribution

Intelligent Robotics Lab, Cal Poly Pomona

Page 23: ARCS Presentation 2008

23

Performance (STA)

Problems: Execution time is exceedingly long. The entire tree is generated irrespective.

Need: Faster execution time. Generate only relevant parts of tree.

Intelligent Robotics Lab, Cal Poly Pomona

Page 24: ARCS Presentation 2008

24

Future Work (STA)

Use heuristics to generate only relevant portions of tree. Iterative Deepening A* (IDA*)

Optimize the current version of the algorithm.

Demonstrate the complete approach using ASyMTRe-D.

Intelligent Robotics Lab, Cal Poly Pomona

Page 25: ARCS Presentation 2008

25

Conclusion (STA)

Multiple tasks are being considered. All possible mapping of tasks to coalitions. No restriction on coalition size.

Reduced overall execution time. STA (Higher level). ASyMTRe-D (Lower level).

An Anytime Algorithm!!

Intelligent Robotics Lab, Cal Poly Pomona

Page 26: ARCS Presentation 2008

26

References Sandholm, T. 1999. Algorithm for Optimal Winner Determination in Combinatorial Auctions,

International Joint Conference on Artificial Intelligence (July). Stockholm, Sweden, 542-547.

Parker, L.E., Tang, F., 2005. ASyMTRe: Automated Synthesis of Multi-Robot Task Solutions through Software Reconfiguration. In Proceedings of IEEE International Conference on Robotics and Automation (April).

Tang, F., Parker, L. E. , 2005. Distributed Multi-Robot Coalitions through ASyMTRe-D. Proceedings of IEEE/RSJ International Conference on Intelligent Robots and Systems (August).

Parker, L.E. and Tang, F., 2006. Building Multirobot Coalitions through Automated Task Solution Synthesis. In Proceedings of the IEEE v. 94 No. 7 (July). 1289-1305.

Parker, L.E., Tang, F., 2007. A Complete Methodology for Generating Multi-Robot Task Solutions using ASyMTRe-D and Market-Based Task Allocation. To Appear in the IEEE International Conference on Robotics and Automation (April).

Wikipedia. http://www.wikipedia.org/

Intelligent Robotics Lab, Cal Poly Pomona

Page 27: ARCS Presentation 2008

27

Questions??Intelligent Robotics Lab, Cal Poly Pomona