Some of us always had a feeling – model based development saves time and money, others of us had reasonable doubts, most of us had endless discussions about it. Good news for all Enterprise Architect and AMUSE users – finally we got proved numbers! The following dissertation shows why it’s true why the benefits definitely overweighs the costs and why MBSE  (Model Based System Engineering) is the approach of the future development.

The only question keeps open – how to use Enterprise Architect and AMUSE to get maximum benefit of these tools, e.g.

  • can I build functional architecture from customer requirements using Enterprise Architect??
  • or can I combine EA and Matlab/Simulink in a reasonable way??

The answer is – yes you can! But it’s a stuff for one of the next articles Winking smile

Summary of the dissertation “Model Based Development of Embedded Software Systems in the Automotive – Costs and Benefits” Author: Sascha Kirstan; Academic Advisor: Prof. Dr. Dr. h.c. Manfred Broy; Technical University Munich (Software & Systems Engineering).
Hardcover Edition:

The dissertation defined and verified following 26 hypotheses – enjoy!

Hypothesis 1

Usage of MBSE saves at least 30% of time during development lifecycle

Fully confirmed
(up to 36%)

Hypothesis 2

Usage of MBSE leads to productivity improvement

Fully confirmed

Hypothesis 3

Usage of MBSE allows cost savings for single projects (opposite of product-lines)

Trend confirmed

Hypothesis 4

The cost-saving effect is only true for product-lines

Statement not possible

Hypothesis 5

The cost-saving effect is bigger for big projects than for small projects

Trend confirmed

Hypothesis 6

The MBSE pay-off is bigger in the domains with high proportion of controlling engineering

Statement rejected

Hypothesis 7

The cost-saving effect is as bigger as more comprehensive the MBSE usage is.

Statement rejected

Hypothesis 8

Usage of MBSE allows saving 20% of maintenance

Trend confirmed

Hypothesis 9

The time-saving effect is bigger than the cost-saving effect

Fully confirmed

Hypothesis 10

– MBSE approach is reasonable for innovative development.
– Classic approach is reasonable for safety-relevant development

Statement rejected

Hypothesis 11

MBSE supports the trend of function-development by OEM’s them self

Fully confirmed

Hypothesis 12

Usage of MBSE based requirement-validation allows finding 10% of all potential errors in the requirement engineering phase

Fully confirmed
(up to ~31%)

Hypothesis 13

Usage of MBSE based architecture-verification allows finding 5% of all potential errors in the architecture engineering phase

Fully confirmed

Hypothesis 14

As higher the formalization-level of the architecture-models as higher the their reusability

Trend confirmed

Hypothesis 15

Usage of MBSE based design-validation increases the number of found errors in the design phase et least by 30%

Fully confirmed
(average 60%)

Hypothesis 16

Reuse of functional models allows saving at least 80% of development costs

Trend confirmed
(average ~50%)

Hypothesis 17

Modeling-level of SW-design has the biggest impact on the development costs

Fully confirmed

Hypothesis 18

Continuity between the SW-design and the implementation allows saving at least 50% of implementation-costs and -time

Trend confirmed
(average ~45%)

Hypothesis 19

Manual implemented code is more efficient than the code generated from a model

Trend confirmed

Hypothesis 20

Model based development allows costs-saving of code-reviews

Statement rejected

Hypothesis 21

Code-generation increases time required for implementation of changes at least by 10%

Statement rejected

Hypothesis 22

Usage of MBSE allows saving 10% of test-time

Fully confirmed
(average ~12%)

Hypothesis 23

Frontloading of test activities (reached due to MBSE usage) decreases number of errors during component-, integration- and system-test

Fully confirmed

Hypothesis 24

Reusability of model-based test-cases is higher than the reusability of classical test-cases

Fully confirmed
(~52% reuse)

Hypothesis 25

Model-based approach saves the costs of test-case-development at least by 25%

Trend confirmed

Hypothesis 26

Model-based approach highly improves the quality asspects, test coverage, automation-level of tests

Trend confirmed