command agents that make human like decisions f

11
COMMAND AGENTS THAT MAKE HUMAN-LIKE DECISIONS FOR NEW TACTICAL SITUATIONS Masood Raza and Dr Venkat V S S Sastry

Upload: eassessment-in-practice-symposium

Post on 28-May-2015

149 views

Category:

Documents


1 download

DESCRIPTION

human like decision making using neural networks, RPD and Soar

TRANSCRIPT

Page 1: Command Agents That make Human Like Decisions F

COMMAND AGENTS THAT MAKE HUMAN-LIKE

DECISIONS FOR NEW TACTICAL SITUATIONS

Masood Raza and Dr Venkat V S S Sastry

Page 2: Command Agents That make Human Like Decisions F

RPD architecture

Page 3: Command Agents That make Human Like Decisions F

Soar Cognitive architecture

Page 4: Command Agents That make Human Like Decisions F

RPD and Soar Cognitive architecture

Page 5: Command Agents That make Human Like Decisions F

RPD-Soar Agent

four inputs Three hidden layers,12 neurons

# known situations

Page 6: Command Agents That make Human Like Decisions F

Experiment – EnvironmentUse NN for recognizing the situation,

not to produce a plan directly

Page 7: Command Agents That make Human Like Decisions F

Experiment - 1

 

Page 8: Command Agents That make Human Like Decisions F

Experiment - 2

The neural net is used to prioritize the strategies according to the recognition value of the situation given to the agent.The RPD-Soar agent is used to reason with the plan. The RPD-Soar agent evlauates the recognized situation and if it is not taking advantage of any of the hills present in the vicinity of the enemy then it discards this strategy and tries the strategy which has the next highest recognition value.

A set of twelve new situations are presented to the agent that sufficiently explores the problem space.

Page 9: Command Agents That make Human Like Decisions F

Analysis

 

Page 10: Command Agents That make Human Like Decisions F

Analysis

• 10 out of 12 situations are correctly recognized

• Situation on the right does not take advantage of the hills, and the Soar-RPD agent picks the next highest matching situation

Page 11: Command Agents That make Human Like Decisions F

Summary

• The neural net successfully recognizes the closest known situation.

• As the strategy is fixed for known situation therefore some times the hills are located in such a way in a new situation that although the fire support or the assault group is very close to the hill but is not able to take tactical advantage from the hill.

• Satisficing seems a good strategy and reduces the search in the problem space.

• Consider re-evaluation of plans, instead of discarding them