Goto

Collaborating Authors

 bucket brigade


Stochastic Market Games

#artificialintelligence

The algorithm works on the basis of entities called classifiers, which are composed of a condition and an action. When the system state matches the condition of an entity, it computes a bid and the entity with the highest bid is allowed to push its specific action onto an execution queue. The bid is distributed among the previously active classifiers that led the system to the triggering condition, such that chains of rewards originate and learning may take place.