Cognition · Software component

Decision Engine

Software componentCognitionCognition & MemoryVariation point (abstract)arc:DecisionEngine

An abstract cognition component that selects which action an agent takes from the currently available options given its model of the current state.

Responsibility. Selects the agent's next action from the set of available options.

Also known as: Action selector, Decision maker, Agent decision-making architecture, Decision-making component, Decision-making paradigm, Decision component, Loan screening agent, Diagnostic pre-screening agent, Legal case prioritization agent, Fraud detection agent

Direct neighbourhood (hover for relationship types)

Variants

Variant	When to choose
Goal-Based Decision Engine	Choose when binary goals suffice and no trade-offs exist; offers faster development and easier debugging.
Heuristic Decision Engine	Choose when real-time decisions must be made in milliseconds and a 'good enough' answer is acceptable, and when the heuristic's systematic bias causes minimal harm in the context (e.g., routine, clear-cut cases).
Hybrid Decision Arbiter	Choose when the problem mixes decision types with incompatible requirements, needs both adaptation and transparency, is safety-critical in an open world, or combines structured knowledge with unstructured perception (high decomposability, hard constraints, mixed data availability).
Learned-Policy Decision Engine	Choose when the strategy space is too large to specify manually, the environment is dynamic or partially unknown, and sufficient training data and computation are available.
MDP Policy Solver	Choose for sequential decisions where current actions affect future options and the state is (or has been augmented to be) Markovian.
POMDP Policy Solver	Choose when the problem is non-Markovian or history-dependent (e.g., treatment response depending on prior treatments) and state augmentation is insufficient.
Pareto Frontier Optimizer	Choose when multiple stakeholders hold diverse or disputed preferences, when trade-offs must be explored before committing, or when objectives are incommensurable and resist a common scale.
Rule-Based Decision Engine	Choose when rules capture the domain completely and consistently produce correct decisions, and transparency and predictability are paramount.
Utility-Based Decision Maker	Choose when decisions involve complex trade-offs among multiple objectives with no clear priority ordering, probabilistic outcomes, or continuous optimization where the degree of success matters more than binary goal achievement; use scalarized weights when a single decision-maker has clear preferences.

Relationships

hosts structural

Predictive Decision Model Ch9.7

is invoked by dependency

is guarded by control

is overridden by control

is audited by assurance

Bias Evaluator Ch10.5

is evaluated by assurance

Evaluation Harness Ch10.5
Preference Annotator abstract Ch9.6

is monitored by assurance

Design guidance

SHOULD choose the decision architecture (goal-based, rule-based, utility-based) by weighing flexibility and optimality against engineering effort, transparency and computational budget.
SHOULD select the decision paradigm (rules, utility, learning, or hybrid) per decision based on explainability, adaptability, optimality and maintainability requirements.
SHOULD combine paradigms in layered architectures, using rules for hard constraints, utility optimisation for continuous trade-offs, and learning for discovering patterns.
SHOULD select the decision paradigm per decision type from strategy-space size, environment dynamics, explainability and verification requirements, and data availability.
SHOULD consider hybrid combinations, using each paradigm where it works best rather than forcing a single paradigm across the whole system.

Classification

Patterns: Agent program / action selectionRule-based decision makingUtility-based decision makingLearning-based decision makingHybrid decision making
Quality attributes: Maintainability (ISO/IEC 25010)Flexibility (ISO/IEC 25010)Explainability (NIST AI RMF: explainable and interpretable)Functional suitability: correctness and validity (ISO/IEC 25010 | NIST AI RMF: valid)

Sources

Ch5.10: T. Nguyen, "Utility-Based Decision Making Fundamentals," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 5.10. ISBN: 9798244538229.
Ch5.11: T. Nguyen, "Rule-Based Decision Making Fundamentals," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 5.11. ISBN: 9798244538229.
Ch5.12: T. Nguyen, "Learning-Based Decision Making Fundamentals," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 5.12. ISBN: 9798244538229.
Ch5.13: T. Nguyen, "Hybrid Decision Systems Fundamentals," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 5.13. ISBN: 9798244538229.
Ch10.5: T. Nguyen, "Human-over-the-Loop," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 10.5. ISBN: 9798244538229.