Red and Yellow robot are initialized with the following goals. Using Linear Social MDP model, at different levels of reasoning, each robot recursively estimates the goals of another by observing their actions.
| Yellow Robot | Red Robot | ||||||
|---|---|---|---|---|---|---|---|
| Social Goal 1 | Social Goal 2 | Physical Goal 1 | Physical Goal 2 | Social Goal 1 | Social Goal 2 | Physical Goal 1 | Physical Goal 2 |
| Compete on Fire | None | Tree | Saw | Exchange on Tree | None on Saw | Fire | Tree |
Level 0 |
Level 1 |
|
|
|
Level 2 |
Level 3 |
|
|
|