Guideline: Model Metrics for Systems Engineering

Introduction

This collection of metrics is based on the Computing Model Complexity whitepaper written by Bruce Douglass. The following section introduces just a subset mostly relevant for the Systems Engineering space. For a full explanation please see the attached document: Computing Model Complexity.pdf.

Model Metrics

Metric Type	Goal	Metric
Model	Org Model	NE - number of distinct model elements
Model	Org Package Density	AEP = Ave elements per package = NE/Number of Packages
Requirements	Complexity	NUCS = Number of UC states (either Total or Per Use Case)
Requirements	Complexity	NIM = Number of unique incoming messages or events (either Total or Per Use Case)
Requirements	Complexity	DCC = Douglass Cyclomatic Complexity for UC state machine
Architecture	Function Size	NUF = Total number of unique functions offered in subsystems interfaces
Architecture	Data Size	NUD= total number of unique semantic data types produced or consumed (either per subsystem or total)
Architecture	Correctness	RNA = Number of requirements not allocated to at least one subsystem
Architecture	Tech complexity	ND = number of engineering disciplines involved in each subsystem (e.g. software, digital electronics, analog electronics, mechanical, pneumatic, biomedical, etc)