Cognition · Software component
Reasoning Strategy Router
Software componentCognitionCognition & Memoryarc:ReasoningStrategyRouter
A routing component that sends each query either to single-path chain-of-thought reasoning or to multi-sample self-consistency according to estimated difficulty, stakes and cost policy.
Responsibility. Selects the reasoning strategy tier for each query.
Also known as: Dynamic reasoning router
Relationships
invokes dependency
reads dependency
routes to dynamic
Design guidance
- SHOULD reserve self-consistency for complex multi-step reasoning tasks with definable correct answers, base accuracy below ~85-90%, and path-specific rather than systematic errors.
- SHOULD NOT apply self-consistency to factual retrieval, single-dominant-path, open-ended generation or summarisation tasks.
Quantitative guidance
As stated by the sources; verify before use.
- General customer-support queries gained only 0.8% accuracy at 5x cost; the feature was removed after a 2-month trial (Ch5.3).
- EPI cost-sensitivity slope m = -361.09 for Self-Consistency vs -177.22 for CoT (Ch5.3).
Classification
- Patterns
- Complexity-based routingReserve expensive reasoning for high-stakes queries
- Quality attributes
- Cost efficiencyPerformance efficiency (ISO/IEC 25010)Functional suitability: correctness and validity (ISO/IEC 25010 | NIST AI RMF: valid)
- Risks mitigated
- Paying multi-sample cost on tasks with no reasoning diversity
Sources
- Ch5.3: T. Nguyen, "Self-Consistency Fundamentals," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 5.3. ISBN: 9798244538229.