Data-driven evaluation of design variants


  • inadequate documentation of the design process, especially subjective decisions
  • no adequate support when searching in model-based reference databases
  • Deficits in the development of the inherent knowledge of references including design development
  • inadequate integration of the potential assessment in the design process

Scientific goals

  • Documentation of qualitative decisions in the evaluation and selection of variants
  • Evaluation of variants based on references and the use of (partial) aspects for further development
  • Evaluation of potentials and performance in design branches of similar references expanded by imperfection (uncertainty, vagueness, inaccuracy)


Work program and methods

WP 1: Specialist vocabulary - qualitative assessments

  • Structuring of technical vocabulary and its (partially) formal mapping

WP 2: Relationships between vocabulary and model content

  • Description of the relationships between technical vocabulary and model parameters including formal presentation and database

WP 3: Interaction concept - recording qualitative evaluation

  • Conception of a text-based episodic description (design stories), selection lists of keywords and "visual keys",
  • Studies on visual representations


WP 4: Potential assessment using references

  • Investigation of activities in the assessment of potential
  • Deriving criteria and identifying parameters

WP 5: Similarity determination and search

  • Investigation of text-based similarity search
  • Definition of text and graph based similarity metrics
  • Definition of global similarity measures

WP 6: Visualization methods of potentials

  • Examination of the representation of potentials and their strengths and weaknesses profiles & the levels of detail including the evaluation in design branches of the references


  • Deriving a concept for recording the evaluation of architectural quality and potential evaluation, e.g. thermal comfort, daylight comfort and structural system (cooperation with WP4 and WP5)
  • Conception of similarity measures (collaboration with WP2 on graph-based retrieval and search methods)
  • Development of a repository, which in addition to semantic building models also includes the design history including evaluations and decisions (cooperation with WP1 and WP2)
  • Prototypical implementation and evaluation (cooperation with WP1, WP2, WP4 and WP5)