Class | Description |
---|---|
DomainEnvironmentWrapper |
If a problem is best described by an Environment class, this domain wrapper takes a domain object
that specifies the state representation and action set, and creates a new domain object in which
actions of the same name and parameterization will make calls to the provided Environment class
and wait for it to return the resulting state before returning themselves.
|
Environment |
In some cases it may be useful to have agents interact with an external environment that handles the current state, execution of actions, and rewards
and maintains other important external information, rather than use the standard Domain, Action, RewardFunction TerminalFunction paradigm of BURLAP.
|