The main class encapsulating the planning algorithm in Grail. Responsible for finding paths in an abstract plans space.
You can use this class directly (as a raw algorithm) or indirectly through a Reasoners.PlannerReasoner instead. More...

Classes
struct	AbstractPlan

class	Config
	The Planner run-time configuration parameters. More...

Public Member Functions
	Planner ()
	Constructs a new planner object with no initial state set. More...

	Planner (WorldState initialState)
	Constructs a new planner object with a provided initial state and a default configuration. More...

	Planner (WorldState initialState, Config config)
	Constructs a new planner object with a provided initial state and the provided run-time configuration. More...

void	PushCondition (ConditionFunction condition)
	Adds a goal satisfaction condition to the planner. The planner may have multiple independent conditions defined. More...

void	IterateWithSnapshot (int iterations, PlannerReasonerSnapshot snapshot, long minDebugSnapshotIteration=0, long maxDebugSnapshotIteration=long.MaxValue)
	Runs the planning process in a DEBUG mode for the given number of iterations. The algorithm will try to formulate a plan starting from the initial state to a state that meets goalConditions. It will call OnPlanningFailed or when finished. More...

void	Iterate (int iterations)
	Runs the planning process for the given number of iterations. The algorithm will try to formulate a plan starting from the initial state to a state that meets goalConditions. It will call OnPlanningFailed or when finished. More...

void	Reset (WorldState initialState)
	Performs a full reset of the planner. A typical usage is before a new decision point that requires a new plan. Please note that this includes setting a new start state, clearing goal conditions, resetting the current iteration number. More...

AbstractPlan	GetPartialPlan ()
	Returns the current best plan, regardless of whether the goal conditions are met. More...

Public Attributes
bool	IsComputing => !isPlanningFinished
	Returns whether the Planner hasn't yet finished computing the plan. More...

Properties
int	IterationLimit `[get, set]`
	Sets or gets the number of iterations per each execution of the planning problem. Each iteration ends with one candidate plan. More...

int	MaxPlanLength `[get, set]`
	Sets of gets the maximum number of actions (steps) a plan can take. Set `-1` for unlimited plans. More...

float	MaxPlanCost `[get, set]`
	Sets or gets an upper bound on the plan cost measured as the sum of costs of actions it consists of. Set `-1` for no limit. It may limit the possible plans generated (if they would exceed this cost). More...

PlannerHeuristic	Heuristic `[set]`
	Enables the planner to heuristically evaluate the calculated plans. Similarly to A* pathfinding heuristic, it should return the estimation of the distance (cost) between an arbitrary state and the goal state. More...

PlanFoundDelegate	OnPlanFound `[set]`
	Callback executed whenever a new plan has been found by the algorithm. More...

System.Action	OnPlanningFailed `[set]`
	Callback executed if no plans were found in the allotted iteration budget and satisfying the MaxPlanLength and MaxPlanCost limits. More...

Detailed Description

The main class encapsulating the planning algorithm in Grail. Responsible for finding paths in an abstract plans space.
You can use this class directly (as a raw algorithm) or indirectly through a Reasoners.PlannerReasoner instead.

Constructor & Destructor Documentation

◆ Planner() [1/3]

Grail.Planners.Planner.Planner ( )

inline

Constructs a new planner object with no initial state set.

◆ Planner() [2/3]

Grail.Planners.Planner.Planner ( WorldState initialState )

inline

Constructs a new planner object with a provided initial state and a default configuration.

Parameters

initialState The initial state of the planning space. The plans will be created from this state to the goal state.

◆ Planner() [3/3]

Grail.Planners.Planner.Planner	(	WorldState	initialState,
		Config	config
	)

inline

Constructs a new planner object with a provided initial state and the provided run-time configuration.

Parameters

initialState	The initial state of the planning space. The plans will be created from this state to the goal state.
config	Configuration with parameters for the planner.

Member Function Documentation

◆ GetPartialPlan()

AbstractPlan Grail.Planners.Planner.GetPartialPlan ( )

inline

Returns the current best plan, regardless of whether the goal conditions are met.

Returns: The current best plan or an empty plan if not enough calculations were performed.

◆ Iterate()

void Grail.Planners.Planner.Iterate ( int iterations )

inline

Runs the planning process for the given number of iterations.
The algorithm will try to formulate a plan starting from the initial state to a state that meets goalConditions. It will call OnPlanningFailed or when finished.

Parameters

iterations The max number of plan iterations to calculate.

◆ IterateWithSnapshot()

void Grail.Planners.Planner.IterateWithSnapshot	(	int	iterations,
		PlannerReasonerSnapshot	snapshot,
		long	minDebugSnapshotIteration = `0`,
		long	maxDebugSnapshotIteration = `long.MaxValue`
	)

inline

Runs the planning process in a DEBUG mode for the given number of iterations.
The algorithm will try to formulate a plan starting from the initial state to a state that meets goalConditions. It will call OnPlanningFailed or when finished.

Parameters

iterations	The max number of plan iterations to calculate.
snapshot	Debug data is outputted to the snapshot every iteration within min and max snapshot iteration parameters.
minDebugSnapshotIteration	Starting from this iteration number, debug data will be included in the snapshot.
maxDebugSnapshotIteration	The last number of iteration from which the debug data will be included in the snapshot.

◆ PushCondition()

void Grail.Planners.Planner.PushCondition ( ConditionFunction condition )

inline

Adds a goal satisfaction condition to the planner.
The planner may have multiple independent conditions defined.

Parameters

condition A boolean function that returns true if the specific goal condition has been reached and false otherwise.

◆ Reset()

void Grail.Planners.Planner.Reset ( WorldState initialState )

inline

Performs a full reset of the planner. A typical usage is before a new decision point that requires a new plan.
Please note that this includes setting a new start state, clearing goal conditions, resetting the current iteration number.

Parameters

initialState The new initial state the formulated plans will start from.

Member Data Documentation

◆ IsComputing

bool Grail.Planners.Planner.IsComputing => !isPlanningFinished

Returns whether the Planner hasn't yet finished computing the plan.

Property Documentation

◆ Heuristic

PlannerHeuristic Grail.Planners.Planner.Heuristic

set

Enables the planner to heuristically evaluate the calculated plans.
Similarly to A* pathfinding heuristic, it should return the estimation of the distance (cost) between an arbitrary state and the goal state.

◆ IterationLimit

int Grail.Planners.Planner.IterationLimit

getset

Sets or gets the number of iterations per each execution of the planning problem. Each iteration ends with one candidate plan.

◆ MaxPlanCost

float Grail.Planners.Planner.MaxPlanCost

getset

Sets or gets an upper bound on the plan cost measured as the sum of costs of actions it consists of. Set -1 for no limit.
It may limit the possible plans generated (if they would exceed this cost).

◆ MaxPlanLength

int Grail.Planners.Planner.MaxPlanLength

getset

Sets of gets the maximum number of actions (steps) a plan can take. Set -1 for unlimited plans.

◆ OnPlanFound

PlanFoundDelegate Grail.Planners.Planner.OnPlanFound

set

Callback executed whenever a new plan has been found by the algorithm.

◆ OnPlanningFailed

System.Action Grail.Planners.Planner.OnPlanningFailed

set

Callback executed if no plans were found in the allotted iteration budget and satisfying the MaxPlanLength and MaxPlanCost limits.

The documentation for this class was generated from the following file:

GrailPlanners/source/Planner/Planner.cs

Classes

Public Member Functions

Public Attributes

Properties

Detailed Description

Constructor & Destructor Documentation

◆ Planner() [1/3]

◆ Planner() [2/3]

◆ Planner() [3/3]

Member Function Documentation

◆ GetPartialPlan()

◆ Iterate()

◆ IterateWithSnapshot()

◆ PushCondition()

◆ Reset()

Member Data Documentation

◆ IsComputing

Property Documentation

◆ Heuristic

◆ IterationLimit

◆ MaxPlanCost

◆ MaxPlanLength

◆ OnPlanFound

◆ OnPlanningFailed