This approach clearly points in the right direction, yet it is based on computer systems as they are. It is still necessary (1) to distinguish between the source of complexity (business objects and logic, systems functionalities, platforms technologies), and (2) the dynamics with environments.
http://caminao.wordpress.com/about/community/requirements-functionpoints/
http://caminao.wordpress.com/2014/06/24/ea-entropy/
http://caminao.wordpress.com/about/community/requirements-functionpoints/
http://caminao.wordpress.com/2014/06/24/ea-entropy/
Article Complexity Reduction: A Pragmatic Approach
Remy:
Excellent question.
One answer to this question is to use well defined processes, like Interpretive Structural Modeling and/or Interactive Management. These processes were developed by Warfield and have been further developed by the System Concepts team.
Systems Concepts has just published two technical reports that address some aspects of your question. These reports are available at:
https://www.researchgate.net/publication/266138960_Structural_Modeling
https://www.researchgate.net/publication/267333858__2013_System_Concepts_LLC_Page_1_of_22_System_Analysis_and_Identification_Objects_Relations_and_Clusters
More detailed aspects of the answer are in work and will be published as technical reports at a later date.
The Structural Modeling paper details one method that may be used to gather empirical information in a very efficient and cost effective manner.
Have fun,
Technical Report Structural Modeling
Technical Report © 2013 System Concepts LLC Page 1 of 22 System Analysis and ...
Remy:
Entropy metrics apply to all types of systems, not just computer based systems.
The source of the complexity is important but is addressed as part of the process.
Lets use the Combs Filter (Union Rule Combination in Fuzzy Logic) as an example.
See:
http://marksmannet.com/RobertMarks/REPRINTS/2004_OnTheUseOfFourierMethods.pdf
http://books.google.com/books?id=SZoYWeKl1aMC&pg=PA307&lpg=PA307&dq=William+Combs+URC&source=bl&ots=rHMjncaPQM&sig=mgD267Tns9F3YflvIQVAnpWC-YY&hl=en&sa=X&ei=O11RVLCYC-_9sAT8q4CABw&ved=0CCoQ6AEwAg#v=onepage&q=William%20Combs%20URC&f=false
Using the Combs Filter a fuzzy rule set with 6 antecedents may be addresses using a set of six rules. A standard fuzzy rule application with six antecedents would take 46656 rules. The complexity reduction provided by the URC method is associated with the 46650 rules that were eliminated. The same system is modeled with six (6) rules, that is fundamental computational complexity reduction.
The same system modeled two different ways.
Have fun,
Joe
The objective is to measure the intrinsic complexity of the problem, not the complexity of the different representations.
Remy:
Measuring the intrinsic complexity of a problem is an important object.
Limiting the scope of complexity reduction to this area only may not be the best approach.
However, complexity metrics with a limited scope are better than no complexity metrics at all.
Have fun,
- Badges
- Science topic
- Similar topics
- Engineering
- Control Theory
- Control Systems
- Control Systems Engineering