INDEX20 RESEARCH · MATHEMATICAL MODEL 02
Crisp, fuzzy and probabilistic sets
A transparent decision model for matching engineering projects with technical capabilities without hiding hard constraints or uncertainty inside one opaque score.
Crisp set
A requirement either passes or fails.
C(r) ∈ {0,1}Examples: required HDL, export restriction, minimum clock, target process availability.Fuzzy set
Suitability can be partial.
μ(r) ∈ [0,1]Examples: domain experience, schedule fit, integration readiness, documentation quality.Probability
Confidence is represented separately.
P(success | evidence)Estimated from evidence quality and comparable completed work; it is not the same as fuzzy membership.DECISION PIPELINE
Keep three questions separate
All mandatory crisp constraints must pass.
Weighted fuzzy membership ranks viable matches.
Probability expresses uncertainty in the available evidence.
A person approves the shortlist and commercial terms.
REFERENCE FORMULATION
Auditable matching score
Fit(p,c) = Σ wj μj(p,c) / Σ wj
Decision = Eligible × Fit × P(success | evidence)
The formula is a research baseline. Production weights, evidence calibration and acceptance thresholds require validation on real projects. No automated decision should replace legal, security or export-control review.
Example inputs for hardware projects
Project requirements
- RTL/Verilog, FPGA or ASIC target
- Memory, GPU/video, edge AI or control block
- Area, timing, power and verification constraints
- Budget, schedule and confidentiality
Capability evidence
- Tools, libraries and process access
- Comparable verified work
- Availability and integration support
- Evidence recency and completeness