The research approach aims to create a tool to assist the building practitioner in finding the best design strategy according to competing objectives. Three core algorithms will be developed, which are based in previous works of the research team. The first algorithm is capable of generating alternative floor plan designs according to the same preferences and requirements of the user. It is a hybrid evolutionary strategy approached enhanced with a local search technique. This algorithm combines several requirements that were otherwise spread in different space planning approaches. It has been tested in several multi-level complex design programs with promising results. The second algorithm explores the thermal performance improvement potential of the generate floor plans, by changing sequentially several geometry variables, such as the building orientation, the openings position and size, the design of the shading mechanisms, the interior wall position, etc.. The third algorithm to be developed aims to find the best renovation strategy considering CS and ES in a cost effective manner. The developed algorithms, however, need to be further developed to fully tackle deep building renovations.
Additionally, two other algorithms will be developed to widen and improve the tool usage efficiency. The first algorithm is a trained Artificial Neural Network (ANN) for building performance assessment. It will be used during optimization processes of the core algorithms, as alternative to the dynamic simulation, which requires long computational runtimes. The dynamic simulation will be used in the tool for detailed performance reports and online training of the ANN. The aim of the second algorithm is to facilitate the building survey process. A 3D point cloud algorithm with a logarithmic proportional objective function will be used to capture and automatically generate the existing building geometry.
These five algorithms will be implemented in a BIM add-on user graphical interface, thus integrating this approach in the architectural workflow. As the usability and accessibility of the tool is an important issue for the aiming practitioners, ergonomic tests for human-interface interactions will be carried out. This information will be feedback to the tool developers. Finally, the tool will be tested in real case scenarios of different levels of building renovation.
Therefore, a BIM add-on user graphical interface will be developed to exhibit the building survey retrieval information, the optimization progress, the different alternatives computed (non-dominated solutions), and detailed performance of each building solution. As the graphical interface has a significant importance to the acceptance of tool by the architects, usability and accessibility tests will be carried out to improve the whole user experience. So, to develop this tool, the following objectives were defined:
- Determine set of requirements;
- Formulate mathematical models;
- Develop the optimization algorithms;
- Develop complementary algorithms;
- Implement a graphical interface;
- Test usability and accessibility;
- Verify and validate the prototype tool;
- Perform real case scenarios;
- Write a tool guidebook; and,
- Organize a technical seminar.