CS344M Autonomous Multiagent Systems Patrick MacAlpine Department - PowerPoint PPT Presentation

CS344M Autonomous Multiagent Systems Patrick MacAlpine Department of Computer Science The University of Texas at Austin

Good Afternoon, Colleagues Are there any questions? Patrick MacAlpine

Good Afternoon, Colleagues Are there any questions? − How does self-interest relate to coordination? Patrick MacAlpine

Good Afternoon, Colleagues Are there any questions? − How does self-interest relate to coordination? − What is special/proprietary about Kiva? Patrick MacAlpine

Good Afternoon, Colleagues Are there any questions? − How does self-interest relate to coordination? − What is special/proprietary about Kiva? − Why not treat resource allocation as an optimization problem? Patrick MacAlpine

Good Afternoon, Colleagues Are there any questions? − How does self-interest relate to coordination? − What is special/proprietary about Kiva? − Why not treat resource allocation as an optimization problem? − What happens if power goes out in AIM? Patrick MacAlpine

Good Afternoon, Colleagues Are there any questions? − How does self-interest relate to coordination? − What is special/proprietary about Kiva? − Why not treat resource allocation as an optimization problem? − What happens if power goes out in AIM? − Further work on AIM? Patrick MacAlpine

Good Afternoon, Colleagues Are there any questions? − How does self-interest relate to coordination? − What is special/proprietary about Kiva? − Why not treat resource allocation as an optimization problem? − What happens if power goes out in AIM? − Further work on AIM? − Can AIM be done without a central controller? Patrick MacAlpine

Good Afternoon, Colleagues Are there any questions? − How does self-interest relate to coordination? − What is special/proprietary about Kiva? − Why not treat resource allocation as an optimization problem? − What happens if power goes out in AIM? − Further work on AIM? − Can AIM be done without a central controller? − Why not just use collison avoidance (Clearpath) for cars? Patrick MacAlpine

Logistics • Surveys — keep them coming Patrick MacAlpine

Logistics • Surveys — keep them coming Patrick MacAlpine

Proposals • Overall, very good! Patrick MacAlpine

Proposals • Overall, very good! • A few common problems: Patrick MacAlpine

Proposals • Overall, very good! • A few common problems: − No clear intro / problem statement Patrick MacAlpine

Proposals • Overall, very good! • A few common problems: − No clear intro / problem statement − Lots of “what” but very little “how”. . . Patrick MacAlpine

Proposals • Overall, very good! • A few common problems: − No clear intro / problem statement − Lots of “what” but very little “how”. . . − . . . or too much how without identifying the challenges Patrick MacAlpine

Proposals • Overall, very good! • A few common problems: − No clear intro / problem statement − Lots of “what” but very little “how”. . . − . . . or too much how without identifying the challenges − No clear description of how you would evaluate success Patrick MacAlpine

Proposals • Overall, very good! • A few common problems: − No clear intro / problem statement − Lots of “what” but very little “how”. . . − . . . or too much how without identifying the challenges − No clear description of how you would evaluate success − Too much proposed Patrick MacAlpine

Proposals • Overall, very good! • A few common problems: − No clear intro / problem statement − Lots of “what” but very little “how”. . . − . . . or too much how without identifying the challenges − No clear description of how you would evaluate success − Too much proposed − Not enough to convince me that it will work Patrick MacAlpine

Proposals • Overall, very good! • A few common problems: − No clear intro / problem statement − Lots of “what” but very little “how”. . . − . . . or too much how without identifying the challenges − No clear description of how you would evaluate success − Too much proposed − Not enough to convince me that it will work − Provide more specific steps/plans Patrick MacAlpine

Proposals • Overall, very good! • A few common problems: − No clear intro / problem statement − Lots of “what” but very little “how”. . . − . . . or too much how without identifying the challenges − No clear description of how you would evaluate success − Too much proposed − Not enough to convince me that it will work − Provide more specific steps/plans − Very little organization- sections Patrick MacAlpine

Proposals • Overall, very good! • A few common problems: − No clear intro / problem statement − Lots of “what” but very little “how”. . . − . . . or too much how without identifying the challenges − No clear description of how you would evaluate success − Too much proposed − Not enough to convince me that it will work − Provide more specific steps/plans − Very little organization- sections − Writing is communication Patrick MacAlpine

− Revise (and visit Carlos at the writing center) • Comments both from me and Sanmit Patrick MacAlpine

Progress reports • Those considering machine learning and using condor... may be time-consuming Patrick MacAlpine

Progress reports • Those considering machine learning and using condor... may be time-consuming • Gather resources early Patrick MacAlpine

Progress reports • Those considering machine learning and using condor... may be time-consuming • Gather resources early − Will provide SCRAM role assignment code Patrick MacAlpine

Progress reports • Those considering machine learning and using condor... may be time-consuming • Gather resources early − Will provide SCRAM role assignment code − Will add scripts for using condor Patrick MacAlpine

Progress reports • Those considering machine learning and using condor... may be time-consuming • Gather resources early − Will provide SCRAM role assignment code − Will add scripts for using condor − Probably add basic comm protocol for 3D Patrick MacAlpine

Progress reports • Those considering machine learning and using condor... may be time-consuming • Gather resources early − Will provide SCRAM role assignment code − Will add scripts for using condor − Probably add basic comm protocol for 3D − Ask! Patrick MacAlpine

Progress reports • Those considering machine learning and using condor... may be time-consuming • Gather resources early − Will provide SCRAM role assignment code − Will add scripts for using condor − Probably add basic comm protocol for 3D − Ask! • Will be stricter on progress reports Patrick MacAlpine

Progress reports • Those considering machine learning and using condor... may be time-consuming • Gather resources early − Will provide SCRAM role assignment code − Will add scripts for using condor − Probably add basic comm protocol for 3D − Ask! • Will be stricter on progress reports − Be more realistic Patrick MacAlpine

Progress reports • Those considering machine learning and using condor... may be time-consuming • Gather resources early − Will provide SCRAM role assignment code − Will add scripts for using condor − Probably add basic comm protocol for 3D − Ask! • Will be stricter on progress reports − Be more realistic − Be much more specific Patrick MacAlpine

Progress reports • Those considering machine learning and using condor... may be time-consuming • Gather resources early − Will provide SCRAM role assignment code − Will add scripts for using condor − Probably add basic comm protocol for 3D − Ask! • Will be stricter on progress reports − Be more realistic − Be much more specific − May reflect side forays Patrick MacAlpine

Progress reports • Those considering machine learning and using condor... may be time-consuming • Gather resources early − Will provide SCRAM role assignment code − Will add scripts for using condor − Probably add basic comm protocol for 3D − Ask! • Will be stricter on progress reports − Be more realistic − Be much more specific − May reflect side forays − Have something implemented and evaluated Patrick MacAlpine

Progress reports • Those considering machine learning and using condor... may be time-consuming • Gather resources early − Will provide SCRAM role assignment code − Will add scripts for using condor − Probably add basic comm protocol for 3D − Ask! • Will be stricter on progress reports − Be more realistic − Be much more specific − May reflect side forays − Have something implemented and evaluated − Return proposals with it - revise/fix based on proposal feedback Patrick MacAlpine

Intersection Management • Kurt’s slides (videos) Patrick MacAlpine

Intersection Management • Kurt’s slides (videos) • Could you use swarms? Patrick MacAlpine

Intersection Management • Kurt’s slides (videos) • Could you use swarms? • What about security aspects? Risks? Patrick MacAlpine

Intersection Management • Kurt’s slides (videos) • Could you use swarms? • What about security aspects? Risks? • What’s new? Patrick MacAlpine