The upcoming changes caused by climate change pose a major challenge on all fronts. In order to weaken the impact, many houses are currently being renovated. The focus lies on reducing the energy demand by increasing isolation and using new heating concepts, the constant change of the environment might also change the requirements of renovation plans. A rise in temperature, for example, might make it necessary to cool down well isolated houses, which would increase energy demand. Also, the SARS-Cov-2 pandemic showed that unforeseeable events can greatly impact every aspect and especially stall construction projects.

Because of this, renovation plans should be planned in design, as well as execution, to be resilient against such changes. This means that changes in the environment have as little influence as possible on the performance of renovation measures and execution.

The aim of this Franco-German project is AI-based resilient planning and design of renovation projects for houses, as well as the planning of renovations and the surveillance of roads using sensor data and AI-assistance.

OFFIS supports this project on planning renovation projects for houses and creates a simulation environment for simulating such projects. Therefore, Data is collected and extracted from the project partners vast experience in the field. With this data, simulation models are deducted. These simulation models are used with Adversarial Resilience Learning to test the resilience of renovation projects and optimize them. This system should allow decision makers to choose renovation plans that avoid delays in the building process and the necessity to renovate again prematurely due to changes in the environment.

• Adversarial Resilience Learning
• Reinforcement Learning
• Simulation
• Data Science