Grail.Simulation Namespace Reference

Classes
interface	ISimulatedActionTranslator
	Translates actions defined in SimulatedGame onto Behaviors used in Grail. More...
interface	ISimulatedGameAction
	Base class for all actions in SimulatedGame. Derive from it for your actions. More...
interface	ISimulatedGameHeuristic
	Interface that encapsulates an algorithm responsible for action-selection inside SimulatedGame (MCTS). This action-selection is performed in certain situations instead of sampling from the whole action space. The goal is to reduce the combinatorial complexity or just provide scripted behavior in certain moments of the game. More...
interface	ISimulatedGameStateTranslator
class	ISimulatedGameUnit
	Base class of a unit in the SimulatedGame framework. A unit represents part of the game-state and peforms actions that affect the game-state. You should not create instances of this class. Use specializations @SimulatedGameThinkingUnit or @SimulatedGameStochasticUnit instead. More...
class	SimulatedGame
	The main interface class for the SimulatedGame model based on the MCTS algorithm. The class supports definition of a game, running simulations and obtaining results. More...
class	SimulatedGameActionMetadata
	This class represents action defined in the SimulatedGames module with additional statistics (score, visits). More...
class	SimulatedGameHelper
	For internal usage only; helps to perform iterations; contains iteration-global variables etc.
class	SimulatedGameNode
class	SimulatedGameObserverForGUI
class	SimulatedGameRuntime
	An interface that is used to terminate simulation in MCTS and set scores to players. Scores are grabbed from here whenever a simulation ends. It is created automatically. More...
class	SimulatedGameStochasticUnit
	A base class of a unit in SimulatedGame that peforms actions according to some probability distribution. It does not represents any players in the game. It should be used if either (1) randomness affects actions available to other players (2) randomness heavily affects the game outcome. Random effects should be applied as actions of SimulatedGameStochasticUnit. More...
class	SimulatedGameThinkingUnit
	A base class of a unit related to a rational/intelligent player. MCTS chooses actions for this kind of unit. See also SimulatedGameStochasticUnit. All units together represent the game state. More...
class	ContinuousInnerDecisionNode
class	DecisionNode
class	DecisionTree
class	LeafDecisionNode
class	NominalInnerDecisionNode
class	C45Algorithm
	This class encapsulates the C4.5 Algorithm used to generate a decision tree (see Grail.Simulation.DecisionTree) based on the dataset provided. More...
class	C45Counter
class	C45IterationParameters
class	C45SingleIterationResult
class	C45ValueDecisionCounter
class	Dataset
	Container for data in form of [consideration_values][decision], where consideration values are [float, float,...,float] and decision is TDecisionType. USe it for training the decision tree. More...
class	DatasetSample
	Represents elements stored in DataSet. They consist of a vector of consideration values and decision made under those considerations. More...
interface	IVectorizer
	Classes implementing this interface provide training data in the offline learning process. More...
class	OfflineLearner
	This is the main class you will use to peform the offline learning process. Assign it to the unit of a SimulatedGame. It will gather properly vectorized data. You get access to this data by methods of this class. More...
class	TreeDataSample
	Structure for storing results in UniqueTreeDataset. More...
class	HashSetExtensions
class	UniqueTreeDataset
	A class that stores samples, which are values for given considerations (conditions in decision tree) together with statistics of decisions performed under those considerations. When the learning process is stopped, then you can convert this object to a flat Dataset, that picks the most frequently chosen decision for each sample. More...
class	DecisionTreeBinarySerializer
	Used to serialize decision tree in a binary mode. Pass object of this type to the DecisionTree.Serialize method. More...
class	DecisionTreeStringListSerializer
	Used to serialize decision tree in a text mode into array of strings. Pass object of this type to the DecisionTree.Serialize method. More...
class	IDecisionStringSerializer
	A class that defines objects that can serialize decisions (e.g., actions in SimplifiedGame) to a string. More...
class	IDecisionBinarySerializer
	A class that defines objects that can serialize decisions (e.g., actions in SimplifiedGame) to a binary format. More...
class	IDecisionTreeSerializer
	A base class for an object that is passed to Serialize() and Deserialize() methods of DecisionTree. Use the concrete implementations such as @DecisionTreeBinarySerializer or @DecisionTreeStringListSerializer. More...

Enumerations
enum class	DecisionNodeType { NOMINAL_INNER , CONTINUOUS_INNER , LEAF }
enum class	DecisionConsiderationType { NOMINAL = 0 , NUMERIC }
	Defines possible types of considerations (conditions). NOMINAL: represents a switch on values (SWITCH(A): CASE X1...; CASE X2:...; ... CASE XN:...) NUMERIC: represents less than value and greater than value conditions If you need more information, please consult the offline learning manual page. More...

Enumeration Type Documentation

DecisionConsiderationType

enum Grail.Simulation.DecisionConsiderationType

strong

Defines possible types of considerations (conditions). NOMINAL: represents a switch on values (SWITCH(A): CASE X1...; CASE X2:...; ... CASE XN:...) NUMERIC: represents less than value and greater than value conditions If you need more information, please consult the offline learning manual page.