This is the webpage for the UNEDA (Universal Engine for Decision Analysis) software platform. It consists of several code layers, the most central of which are the DTL (Decision Tree Layer) functional API layer and its core calculation library TCL (Tree Core Layer). UNEDA is an open-source platform designed as a general foundation for decision-analytic software. In its basic form, it can handle methods that conform to classic probability and utility theory, but it can be extended to work with any method compatible with additive utilities (such as TOPSIS, VIKOR, or outranking methods). UNEDA brings the following features to any decision-analytic software implementation:
UNEDA was developed between 1994 and 2025, first at the Royal Institute of Technology and later at Stockholm University. A core design principle from the outset was to accommodate imprecision in the input data, ensuring that the platform never demands more precision than the decision-maker can reasonably provide. In practice, input values are seldom known with such certainty that they can be meaningfully represented as exact numbers. The earliest implementations were instances of the framework introduced in the PhD thesis Computational Decision Analysis (link in the left side panel). The framework was subsequently extended to handle decision trees of arbitrary depth, rather than only a single level, and also extended to handle multiple criteria instead of only a single criterion.
Over the years, UNEDA has been used in a variety of projects, both research-focused and commercial. Most project teams built their own layers on top of UNEDA to suit the functional requirements they had for their applications. Two examples of additional layers are bundled together with the basic platform: UNEDA-CAR for cardinal ranking input (otherwise numerical) and UNEDA-SML for a stakeholder group extension. Users are encouraged to develop their own interface layers and, if circumstances permit, publish them for broader access and community benefit.
The UNEDA source code is stored in a repository on GitHub and can be downloaded and used free of charge for any purpose. Background material and documentation are available at this website. In the left side panel, there are links to specifications and documents describing the library package and its algorithms.
UNEDA conforms to the theoretical foundations of prescriptive decision analysis as described in the book Foundations of Computational Decision Analysis (link in the left side panel). Thus, it is compatible with most of the methods described here, viz. all methods built on additive utility. For all compatible methods, UNEDA offers substantial functional extensions in the form of i.a. imprecise input data (numeric and rankings), belief distributions, multiple evaluation modes and sensitivity analyses.
The UNEDA software is licensed under Creative Commons CC BY 4.0. It is provided "as is", without warranty of any kind, express or implied. Reuse and modifications are encouraged, with proper attribution.
