bAction

Baction is a conceptual construct used in planning and robotics to denote a formal representation of an action that a system can perform. It emphasizes the structure and semantics of actions rather than treating them as opaque labels, and it is encountered in discussions of action modeling, decision making, and autonomous behavior.

A typical Baction specifies preconditions, which are the conditions that must hold for the action to be executable, and effects, which describe how the world changes after the action is performed. Many formulations also include a cost or probability parameter to capture resource use, time, or the likelihood of successful execution. Some models extend the definition with duration, observability, and stochastic outcomes.

Variants of Bactions include deterministic forms, where outcomes are certain, and probabilistic forms, which assign outcome distributions. Hierarchical Bactions compose smaller actions into higher-level tasks, aiding scalable planning. In some approaches, the policy for when to trigger a Baction is separated from the action’s internal model of effects, enabling modular reasoning.

Typical applications lie in automated planning, robotic manipulation, and simulation environments, where Bactions provide a common interface between planners and the environment. They support reasoning about capability constraints, feasibility, and resource management in complex tasks.

Example: a Baction called GraspBlock might have preconditions such as the gripper being open and the block being within reach; effects like the gripper being closed and the block being held; and a cost reflecting the time or energy required. In probabilistic models, a success probability might be specified, with alternative states for failed execution.

See also: Action (AI), Planning, PDDL, Robotic manipulation, Action model.