Next: , Previous: , Up: Adaptive Probabilistic Mapping   [Contents][Index]


2.7 Specifying Output Signal Weights

The weight of an output signal of an actor is a coefficient for multiplying the probability of selecting this output signal by the actor. The coefficient is a hint—it increases or decreases that probability. To disable emitting an output signal, set its weight equal to 0.

In the simplest case, the actor uses the weight of an output signal when selecting the output signal in all action choice states. On binding the weight to a specific action choice state, the actor uses that weight when selecting the output signal only in this action choice state.

Large actors do not support setting the weight of an output signal for all action choice states at once. However, all actors support binding output signal weights to specific action choice states.

A preloaded probability profile is a prepared list of output signal weights loaded into an actor. You can assign the preloaded probability profile to specific action choice states for setting output signal weights for those action choice states.

For a small actor, you can also bind output signal weights to an action choice state by specifying a probability profile for the action choice state held in the statistics storage of this small actor. See Structures for Accessing Storage, for how to pass profile probabilities to storage access functions to write them to statistics storage.

It is essential to note that by reasons described in Number of Output Signals, changing the default weights of output signals of a small actor, in general, makes its behavior ill-defined. Therefore, you should avoid doing that. Changing the weights seems to be a correct procedure only for a large actor, because this procedure does not alter the number of choice alternatives at decision points along the pathways to output signals.


Next: , Previous: , Up: Adaptive Probabilistic Mapping   [Contents][Index]