Cognition · Software component
Heuristic Estimator
Software componentCognitionCognition & MemoryVariation point (abstract)arc:HeuristicEstimator
A cognition component that estimates, for a search node, the remaining cost h(n) to the goal, guiding which candidates a search planner expands first.
Responsibility. Estimates remaining cost-to-goal for search nodes.
Also known as: Domain heuristic, Heuristic function, h(n)
Variants
| Variant | When to choose |
|---|---|
| Geometric Distance Heuristic | Choose when state coordinates are available and per-node evaluation must be near-free; the sweet spot for grid and road pathfinding. |
| Learned Heuristic Estimator | Choose only when inference is cheap relative to node expansion cost, so higher accuracy reduces total search time. |
Relationships
is invoked by dependency
reads dependency
Design guidance
- MUST NOT overestimate true remaining cost when the planner's optimality guarantee is relied upon.
- SHOULD choose the most dominant admissible heuristic that is affordable to compute; taking the maximum of several admissible relaxations yields a dominant heuristic.
- SHOULD be judged on total time T_total = T_heuristic x N_expanded + T_overhead, not on node expansions alone.
Classification
- Patterns
- Heuristic rollout guidanceHeuristic node-value initialisationAdmissibilityConsistencyHeuristic dominanceMax of relaxed-problem heuristicsLandmark distance precomputation
- Quality attributes
- Performance efficiency (ISO/IEC 25010)Functional suitability: correctness and validity (ISO/IEC 25010 | NIST AI RMF: valid)
- Risks mitigated
- Loss of optimality from overestimating heuristics
Sources
- Ch5.5: T. Nguyen, "Monte Carlo Tree Search Fundamentals," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 5.5. ISBN: 9798244538229.
- Ch5.6: T. Nguyen, "A* Search and Replaning," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 5.6. ISBN: 9798244538229.