Optimization of daylight in architectural and urban projects

The aim of this communication is to present a formalization of a set of problems involved in the design of buildings or urban settlements with respect to daylight. The complexity of the problem is due to the necessity to take into account the environment of the analyzed objects. Local configuration and presence of other buildings have a strong influence on the possible solutions. They often compel the architect to solve multi scale problems. For simple evaluations, we developed software able to compute quantities of interest at particular points and for specified periods of time. Because we want these quantities being useful for design, we have to make sure that they correspond to the practices of the architects. Moreover when the designer is gathering together the relevant information to make the good choices, he has to analyze a lot of configurations and this process can be long and tedious. So we propose to use optimization in order to provide indications of the best solutions. The first problem is to define an objective function. Usually, it corresponds to the previous quantities of interest, i.e. maximum natural lighting during the whole year but at the same time, a maximum heating in winter and a minimum one in summer. At this point we yet observe that we have to deal with a multi objective problem. The second point is related to the formulation of the constraints: i.e. area or position of the windows, cost of the devices, limitation of the visual impact, selection of standard devices to modify the illumination, etc… We see now that sometimes we need to solve discrete problems. Finally we have to identify the design parameters that correspond to the input of the above analysis program. Most of these parameters are geometric, but sometimes they are elements of catalogs or they have to obey to norms. To make sure that this optimization will meet the requirements of the designers, it is necessary to analyze carefully how architects are performing their design and what kind of information they need in order to take the right decisions.