An adaptive probabilistic mapping2, called an actor for short, interacts with an application program or the environment by the means of exchanging signals. The actor emits output signals3 on the basis of input signals and spur increments supplied to it. A list of occurrences of input and output signals, where for every occurrence there is specified a particular moment of time when that occurrence took place, represents the event history of an actor.
The choice of output signals an actor emits has to be optimal in some sense defined by the developer. Assuming input signals received by an actor are hard to predict, supplying reasonable spur increments is a means of achieving optimality of the choice of output signals emitted. Each spur increment is an increment of the spur of a specific type. See Spur-driven Behavior, for definitions of spur and spur type.
Every spur type is defined for an actor by a way of spur perception, spur weight, and a type of time used to compute spur increment velocity. A list of definitions of spur types along with a method of performing spur increments specify the spur scheme of an actor.
|• Event History Example|
|• Small and Large Actors|
|• Creating an Actor|
|• Incrementing Time|
|• Incrementing Spur|
|• Event History N-gram|
|• Generating an Optimal Action|
|• Customizing the Relative Probability Function|
|• Specifying Weights of Output Signals|
|• Automatic Spur|
|• Controlling Random Behavior of an Actor|
|• Other Parameters of an Actor|
|• Example of Using the Actor API|
Before QSMM version 1.17, the term “optimal action generation engine” was commonly used instead of the term “adaptive probabilistic mapping”.
Before QSMM version 1.17, the term “action signal” was used as a synonym for the term “output signal”. Now the synonym is deprecated.