▎Loyal Wingman

Utility AI

Decision pattern that scores candidate actions by weighted utility curves.

Definition

Utility AI evaluates each candidate action by composing a small set of normalized 'consideration' curves (threat, fuel, time, opportunity) into a scalar utility. The agent then samples or argmaxes across actions. The pattern produces smooth, designer-tunable behavior and is increasingly used to break ties inside behavior trees on autonomous wingmen.

Reference attributes

Output: Scalar utility per action
Strength: Smooth, tunable, explainable

Related terms

Behavior Trees
Hierarchical control structure for composing autonomous agent behaviors.
Goal-Oriented Action Planning (GOAP)
Lightweight STRIPS-style planner that chains actions by precondition and effect.

#decision#autonomy