Problems

The syntax for problem definitions has been extended to allow for the specification of a probability distribution over initial states, and also to permit the association of a one-time reward with entering a goal state. It is otherwise identical to the syntax for PDDL2.1 problem definitions.

$\langle$ problem$\rangle$		::=		( define ( problem $\langle$ name$\rangle$ )
				( :domain $\langle$ name$\rangle$ )
				[$\langle$ require-def$\rangle$]
				[$\langle$ objects-def$\rangle$]
				[$\langle$ init$\rangle$]
				$\langle$ goal$\rangle$ )
$\langle$ objects-def$\rangle$		::=		( :objects $\langle$ typed list (name)$\rangle$ )
$\langle$ init$\rangle$		::=		( :init $\langle$ init-el$\rangle$* )
$\langle$ init-el$\rangle$		::=		$\langle$ p-init-el$\rangle$
		|		:probabilistic-effects ( probabilistic $\langle$ prob-init-el$\rangle$* )
$\langle$ p-init-el$\rangle$		::=		$\langle$ atomic formula (name)$\rangle$  | :fluents ( = $\langle$ f-head (name)$\rangle$ $\langle$ number$\rangle$ )
$\langle$ prob-init-el$\rangle$		::=		$\langle$ probability$\rangle$ $\langle$ a-init-el$\rangle$
$\langle$ a-init-el$\rangle$		::=		$\langle$ p-init-el$\rangle$  |  ( and $\langle$ p-init-el$\rangle$* )
$\langle$ goal$\rangle$		::=		$\langle$ goal-spec$\rangle$ [$\langle$ metric-spec$\rangle$]  |  $\langle$ metric-spec$\rangle$
$\langle$ goal-spec$\rangle$		::=		( :goal $\langle$ GD$\rangle$ ) [( :goal-reward $\langle$ ground-f-exp$\rangle$ )] :rewards
$\langle$ metric-spec$\rangle$		::=		( :metric $\langle$ optimization$\rangle$ $\langle$ ground-f-exp$\rangle$ )
$\langle$ optimization$\rangle$		::=		minimize  |  maximize
$\langle$ ground-f-exp$\rangle$		::=		$\langle$ number$\rangle$  |  $\langle$ f-head (name)$\rangle$
		|		( $\langle$ binary-op$\rangle$ $\langle$ ground-f-exp$\rangle$ $\langle$ ground-f-exp$\rangle$ )
		|		( - $\langle$ ground-f-exp$\rangle$ )
		|		( total-time )  |  total-time
		|		( goal-achieved )  |  goal-achieved
		|		:rewards $\langle$ reward fluent$\rangle$