
Research Update: This research has developed a method called the BIM-Centric Daylight Profiler for Simulation, or BDP4SIM. The general concept of the BDP4SIM method is to step through each space in a building to determine space with “similar” daylighting conditions such that the results from one space can be accurately applied to another “similar” space. The method analyzes each space sequentially, creating a profile of critical daylighting information as it goes and checking each new space against the growing database of daylighting profiles.
In order to properly describe the daylighting conditions of space, along with general building information relative to site location, climate and orientation that is common to all spaces in a given building, the following additional daylighting profile information is often required:
The methodology for gathering this daylighting profile information is described in the following subsections in the order that they are checked. If a space does not pass one of the checks, further checks are not performed to improve simulation time although all the checks discussed only take a fraction of the simulation time. The comparisons are performed in this order:
A number of similarity tolerances or “dials” have been implemented in the software when gathering daylight profile information. This allows a user to adjust the amount of similarity between grouped “similar” spaces. The different tolerance parameters implemented are listed below and are described throughout the different daylight profile sections that follow.
Once similar spaces are matched up, the first found space is simulated and the results are then applied, weighted according to the areas of the similar spaces, to all other similar spaces and used in the building daylight metric totals.

Fischer, M. Iaccarino, G., Welle, B., and Druzgalski, C. (2011). "Improving the Cost-Effectiveness and Scalability of Multidisciplinary Design Optimization (MDO) for Daylighting Simulation using Artificial Intelligence, Distributed Computing, and Uncertainty Analysis." CIFE SEED Proposal 2011.
Problem: The cost-effectiveness and scalability of an MDO process for today’s large and complex high-performance building projects is limited by excessive computational requirements for sustainable analyses, such as daylighting simulation.
Proposed
Research: We
propose to develop
three related methodologies to
drastically improve the scalability and cost-effectiveness of daylighting MDO: (1) develop an
artificial intelligence-based methodology/knowledge-based system for
daylighting design; (2) a product decomposition methodology for parallel and
distributed computing; and (3) a methodology for optimization uncertainty
analysis.


