Economic Goal Selection

From OpenCog

Goals are not special objects within OpenCogPrime; rather, any Atom may be taken as a Goal.

There is a Pool of Supergoals, which are continually allocated STICurrency and LTICurrency merely for being supergoals.

For each Atom G two aspects may be estimated (often inferentially; see PLN):

  • the degree to which G implies the supergoals
  • the feasibility of achieving G (i.e. to what degree G may be achieved in what contexts given what amounts of effort)

The supergoals pass their currency along to other Atoms G based on these two aspects.

And, based on the amount of currency passed to G by the supergoals,

  • the GoalBasedSchemaLearning MindAgent may nominate G as a "goal" for the purpose of triggering a Procedure Learning process aimed at figuring out Schema that are likely to lead to the achievement of G.
  • the GoalBasedSchemaActivation MindAgent may nominate G as a "goal" for the purpose of placing into the Active Schema Pool some Schemata that are estimated likely to lead to the achievement of G

(Note: a bunch of complexity is being swept under the rug here, intentionally; and the above description is in some ways slightly misleading. Currency itself is not directly passed from supergoals to subgoals, but rather more complicated sorts of financial instruments are. The above description is intended just to give the overall flavor of what happens, rather than to actually tell you how to implement the appropriate equations.)