The aim of the research is to develop a digital construction control center. This will allow for transparent statements about the current state and future development of complex construction projects at any time. The need for management measures can be assessed almost in real-time. To achieve this, system dynamic models will be developed together with project managers on selected complex, large scale construction projects. These models will be developed in workshops based on the knowledge and experience of the involved individuals. Knowledge gained through personal observation can complement the model development, as well as available “hard” and “soft” metrics. The models will be calibrated and validated using historical data. The information progressively stored in conventional management tools will be linked with the model to represent the current state of the overall system in near-real-time.

Currently, a variety of different tools is used to manage complex construction projects. Each is tailored to the specific needs of the respective disciplines and serves as a valuable repository of information. This information is periodically compiled and evaluated in a project report. The evaluation of information corresponds to a transformation of disposable knowledge into orientational knowledge and follows the mental models of the project managers. A mental model is the implicit understanding of how project success is achieved. Mental models can be incorrect and incomplete. They are particularly difficult for others to comprehend. Intuition is difficult to articulate. Between two reporting dates, management is in “instrument flight”. Individual metrics can be read, but a clear view of the big picture remains obscured.

System dynamic models can process quantitative information of all formats (management tools) as well as qualitative and descriptive information (mental models). The number of features that can be modeled and the feedback relationships between them is unlimited. Their dynamic behavior over time can be very well simulated using computer support. Based on the current state of the system, forecasts can be continuously made about its future development. The evaluation of information will henceforth be partially automated using the developed project models. In addition, the effects of interventions can be simulated in various scenarios, allowing to derive actionable recommendations.

In the research work construction project data from the initiation, planning and execution to the handover of the object are recorded and systematically analyzed. On this basis, a concept of multi-project management is to be developed in the perspective of an industrial building owner, which enables a data-based and thus objective as well as comparable recording of the project status. Within this framework, the research work deals with the availability, structuring and processing of data in digital systems. Based on this, statistical methods and machine learning approaches are used to investigate re-occurring dynamics and data patterns in construction projects, which lead to the implementation of an early warning system, that can be used for effective control by the owner in multi-project management.