Problem Statement: Implicit Decision Logic
In many public-sector organizations, the underlying decision logic of administrative procedures is not explicitly documented. Decisions about benefit entitlements—for example in the context of housing allowances—typically rely on a wide range of factors, including income level, household size, or rent level. These decision rules are often dispersed across heterogeneous and non-standardized sources of knowledge, such as administrative directives, internal manuals, or the informal experience of case workers.
This implicit distribution of knowledge leads to inconsistencies in decision-making, complicates the onboarding of new staff, and increases the effort required to implement legislative changes. From an information systems perspective, this constitutes a significant organizational risk, particularly with regard to scalability, traceability, and standardization (Janssen & Kuk, 2017).
Conceptual Framing of Decision Logic
In the following, decision logic is understood as the rule-based transformation of input variables into well-defined output decisions within a given decision space. It represents a core component of operational information systems and is closely related to concepts such as business rules and operational decision management (Dumas et al., 2018).
Decision Model and Notation (DMN)
Decision Model and Notation (DMN) is a standard by the Object Management Group for the formal modeling and execution of decision logic (OMG, 2023). While BPMN primarily describes the flow of business processes, DMN focuses on the structured and machine-readable representation of decisions within those processes.
A DMN model typically comprises the following elements:
- Decision Requirements Diagram (DRD): A visualization of decision dependencies and information flows
- Decision tables: A tabular representation of decision rules mapping inputs to outputs
- FEEL (Friendly Enough Expression Language): An expression language for defining conditions and calculations
Example: Simplified Housing Allowance Decision as a DMN Table
The following simplified decision table illustrates how input parameters are mapped to a benefit decision. In a DMN context, this table would be interpreted as an executable decision artifact.
| Income group | Household size | Rent level | Housing allowance |
|---|---|---|---|
| < €1,000 | 1 person | < €400 | €180 |
| < €1,000 | 1 person | €400–600 | €220 |
| €1,000–1,500 | 2 people | < €600 | €150 |
| > €1,500 | – | – | €0 |
The formal specification of such rules can be expressed using FEEL, for example by interval notations like [0..1000] to define income ranges (OMG, 2023).
Integration with BPMN
DMN and BPMN are complementary standards. BPMN models the control flow of a process, while DMN formalizes the decision logic at specific decision points. In BPMN, integration typically occurs via so-called business rule tasks that reference external DMN decision models (Dumas et al., 2018).
This separation enables a clear modularization of process logic and decision logic and provides a foundation for flexible and maintainable information system architectures.
Use Cases in Public Administration
- Eligibility assessments for social benefits based on legally defined thresholds
- Standardized building permit procedures in clearly rule-governed cases
- Fee calculation based on defined service types and effort categories
- Automated routing of applications based on jurisdiction and responsibility rules
- Combining rule-based systems with machine learning for decision support
Implications for Legal Certainty and AI Governance
The explicit modeling of decision logic has implications not only for efficiency but also for governance. First, it increases the transparency of administrative decisions and enables their systematic review, which is a key prerequisite for legal certainty. Second, DMN can serve as a transparent rule base that complements data-driven models and thereby contributes to the explainability of AI systems, particularly in regulated contexts (European Union, 2024). Third, standardized decision logic supports the consistent application of rules and thus helps to ensure equal treatment of comparable cases.
Conclusion
Implicit decision logic represents a structural risk for organizations, especially in highly regulated domains such as public administration. DMN offers a formalized approach to externalizing, standardizing, and automating such logic. From an information systems perspective, DMN can therefore be understood as a key design element of decision-centric information systems.
References
- Dumas, M., La Rosa, M., Mendling, J., & Reijers, H. A. (2018). Fundamentals of Business Process Management (2nd ed.). Springer. https://doi.org/10.1007/978-3-662-56509-4 [web:9]
- Janssen, M., & Kuk, G. (2017). The challenges and limits of big data algorithms in technocratic governance. Government Information Quarterly, 34(3), 371–377. https://doi.org/10.1016/j.giq.2017.08.011 [web:9]
- Object Management Group. (2023). Decision Model and Notation (DMN), Version 1.5. https://www.omg.org/spec/DMN/1.5/ [web:1][web:5]
- European Union. (2024). Regulation (EU) 2024/… on Artificial Intelligence (AI Act). https://eur-lex.europa.eu/ [web:3][web:8]