Decisional DNA and SOEKS: a review of reusable decision knowledge

On this page

1. What the source says
2. Why it matters
3. My perspective
4. Caveats

What if the decisions an organisation has already made could be stored in a form that software can search, compare and reuse when a similar problem comes up? That is the idea behind the Set of Experience Knowledge Structure (SOEKS) and Decisional DNA (DDNA). In a review chapter published by Springer in 2023, A B M Mehedi Hasan and Cesar Sanin (Australian Institute of Higher Education, Sydney) and Md. Nafiz Ishtiaque Mahee (BRAC University, Dhaka) trace how these two concepts have developed since they were first proposed.

The chapter, “Set of Experience Knowledge Structure (SOEKS) and Decisional DNA (DDNA): A Review”, appears in Innovative Technologies in Intelligent Systems and Industrial Applications. This is the proceedings of the CITISIA 2022 conference, published as volume 1029 of Lecture Notes in Electrical Engineering (pp. 925-935).

What the source says

The review starts from a clear purpose. SOEKS and DDNA are designed to capture, store, organise and reuse formal decisions in an explicit knowledge form, so that past experience can support future decision-making.

An earlier description by Cesar Sanin and Edward Szczerbicki shows what this means in practice. In a 2020 chapter they explain that SOEKS draws together filtered information from formal decision events, and that “SOEKS comprises variables, functions, constraints and rules associated in a DNA shape”. They also say it can deal with uncertain and incomplete information. DDNA builds on this by grouping sets of experience into classes called decisional chromosomes. Together these form what the authors call an enterprise’s decisional fingerprint.

The 2023 review has three parts:

• The original concept. It explains SOEKS and DDNA, compares DDNA with human DNA and describes how a DDNA is constructed.
• Advancements over time. It looks at how the concept has been extended, in chronological order.
• Applications and limitations. It discusses where DDNA has been applied and its limitations, and ends by suggesting possible future advancements.

The reference list shows how much work the review covers. Early papers implemented SOEKS in XML (2005 and 2007) and as a Java class within an ontology system (2007). Others proposed heterogeneous similarity metrics for knowledge administration (2006), a negotiation process for dissimilar sets of experience (2007) and genetic algorithms for solving sets of experience (2007), drawing on the wider literature on multiobjective evolutionary algorithms. Later work covered reflexive ontologies and decisional trust (2008), DDNA in embedded systems and robotics (2010), web data mining (2011) and a proposed knowledge market (2013).

More recent work moves into industrial settings. Examples include Virtual Engineering Objects and Virtual Engineering Processes as a form of cyber-physical system (2015), a Virtual Engineering Factory (2016), experience-based knowledge representation for the Internet of Things and cyber-physical systems (2019) and DDNA-based machine monitoring and total productive maintenance within an Industry 4.0 framework (2022). The chapter’s keywords follow the same path: knowledge engineering, evolutionary and genetic algorithms, Industry 4.0 and the Internet of Things.

Why it matters

The central idea is that experience is most useful to a system when it is explicit. SOEKS stores decision knowledge as variables, functions, constraints and rules. In principle, that knowledge can therefore be examined and compared as well as reused.

A consolidated review adds value in three ways.

First, it maps a scattered literature. Work on SOEKS and DDNA spans nearly two decades and many separate papers, a large share of them in the journal Cybernetics and Systems. A chronological account gives newcomers a route through it.

Second, it shows how an idea about representing knowledge moved into applied settings. The progression from XML and Java implementations to robotics, embedded systems, web data mining and smart manufacturing is a case study in how a research concept develops.

Third, it brings an earlier overview up to date. Shafiq, Sanin and Szczerbicki reviewed the past, present and future of SOEKS and DDNA in Cybernetics and Systems in 2014. The 2023 chapter cites that review and also covers later work, including the Internet of Things and Industry 4.0 studies published between 2015 and 2022.

My perspective

In our chapter, Md. Nafiz Ishtiaque Mahee, Cesar Sanin and I set out to review past work and advancements in SOEKS and DDNA since the original concept was introduced. We began with the original idea, including the comparison with human DNA and the construction of DDNA. We then followed its advancements in chronological order before discussing applications, limitations and possible future directions.

Rather than draw firm conclusions here, I want to note some questions this work raises for neighbouring fields.

Ontologies appear throughout the DDNA story, from an early Java implementation within an ontology system to reflexive ontologies. This raises a question for cybersecurity ontology and interoperability. Could decisions made during security work, such as how an incident was assessed and handled, be captured as explicit and reusable experience? And would a shared structure help that experience move between tools and organisations?

There is also a question for agentic AI and automation. Could an explicit store of past decisions, with its variables, constraints and rules visible, help people trace the reasoning behind an automated agent’s recommendation?

For computing education, the chronology may also be useful as a teaching example. It follows a single idea from a conceptual structure to implementations in XML and Java, then to optimisation with genetic algorithms and finally to applications in manufacturing. This raises a question for course design: would following one concept over time help students understand knowledge representation and decision support better than studying each technique on its own?

Caveats

• This post summarises the chapter at the level of its published abstract, keywords and reference list. The chapter itself discusses the specific limitations and proposed future advancements, so readers who want the detail should go there.
• A review brings together earlier work. It does not report new experiments or performance results.
• Most of the 30 works cited are by Cesar Sanin, Edward Szczerbicki and their collaborators, and Cesar Sanin is also a co-author of this review. That makes the chapter a detailed map of this line of research, but readers who want independent comparisons with other decision support approaches will need to look further.

Sources and further reading

1. Set of Experience Knowledge Structure (SOEKS) and Decisional DNA (DDNA): A Review, A. B. M. Mehedi Hasan, Md. Nafiz Ishtiaque Mahee and Cesar Sanin. Innovative Technologies in Intelligent Systems and Industrial Applications (CITISIA 2022), Lecture Notes in Electrical Engineering, vol. 1029, Springer, Cham, 2023, pp. 925-935
2. Set of Experience and Decisional DNA: Experience-Based Knowledge Structures, Cesar Sanin and Edward Szczerbicki. Knowledge Management and Engineering with Decisional DNA, Intelligent Systems Reference Library, vol. 183, Springer, Cham, 2020, pp. 1-48
3. Set of Experience Knowledge Structure (SOEKS) and Decisional DNA (DDNA): Past, Present and Future, S. I. Shafiq, C. Sanin and E. Szczerbicki. Cybernetics and Systems, 45(2), Taylor and Francis, 2014, pp. 200-215