Package jline.lang

Class OutputStrategy

java.lang.Object
jline.lang.OutputStrategy
All Implemented Interfaces:
Serializable
public class OutputStrategy extends Object implements Serializable
Class modelling the output section of a Node
See Also:

  • Field Details

    • legalStrategies

      public static List<RoutingStrategy> legalStrategies
      Routing strategies an OutputStrategy may carry. RL belongs here: it is a state-dependent routing strategy on the same footing as JSQ and KCHOICES (see Network.refreshRouting, Network.sub_rl, SnHasSDRouting and the RoutingStrategy_RL feature), and Node.setRLRouting installs it through this class. Omitting it made setRLRouting throw, so RL routing could not be configured at all. FIRING is excluded: a Transition's outgoing arcs are described by its firing outcomes, not by an output strategy.

  • Constructor Details

  • Method Details

    • getDestination

      public Node getDestination()

    • setDestination

      public void setDestination(Node destination)

    • getJobClass

      public JobClass getJobClass()

    • getProbability

      public double getProbability()

    • setProbability

      public void setProbability(double probability)

    • getRoutingStrategy

      public RoutingStrategy getRoutingStrategy()

    • setRoutingStrategy

      public void setRoutingStrategy(RoutingStrategy routingStrategy)

    • getKchoicesK

      public int getKchoicesK()

    • setKchoicesK

      public void setKchoicesK(int k)

    • isKchoicesWithMemory

      public boolean isKchoicesWithMemory()

    • getRlValueFunction

      public Matrix getRlValueFunction()
      Returns the RL value function, or null when RL routing is not configured.
      Returns:
      the value function

    • getRlValueFunctionShape

      public int[] getRlValueFunctionShape()
      Returns the per-axis shape of the tabular RL value function.
      Returns:
      the shape, or null when not configured

    • getRlNodesNeedAction

      public int[] getRlNodesNeedAction()
      Returns the node indices that consult the RL value function.
      Returns:
      the node indices, or null when not configured

    • getRlStateSize

      public int getRlStateSize()
      Returns the RL state-size hint.
      Returns:
      0 for tabular, >0 for linear approximation, <0 for JSQ fallback

    • setRlParams

      public void setRlParams(Matrix valueFunction, int[] vfShape, int[] nodesNeedAction, int stateSize)
      Stores the RL routing parameters on this strategy.
      Parameters:
      valueFunction - the value function
      vfShape - per-axis sizes of the tabular value function
      nodesNeedAction - node indices that consult the value function
      stateSize - 0 = tabular, >0 = linear approximation, <0 = JSQ fallback

    • setKchoicesWithMemory

      public void setKchoicesWithMemory(boolean withMemory)