Abstract
We developed a method for optimizing lighting design using the genetic algorithm. The floor illuminance of a room was obtained by superimposing the distributions that each light source produced. The illuminance distribution of each light source was calculated using the Monte Carlo Simulation and was stored in a database. The desirable illuminance distribution is obtained by selecting the light sources from the database by genetic algorithm. An optimization condition is represented by the fitness function of a genetic algorithm. To obtain a uniform illuminance distribution, we used the standard deviation of the floor illuminance. The standard deviation of the floor illuminance can be used as an index that represents the uniformity of the illuminance distribution. The lighting design in a simple cubic room is shown as an example. The illuminace distribution yielded by the location of the light sources that were selected using this technique was compared with that of evenly-spaced light sources on the ceiling plane. Although the room that we used was simple, the illuminace distribution obtained by using the technique almost corresponded to the target condition. This technique is practical for optimizing the arrangement of light sources inside a room. Moreover, this technique has various applications such as daylighting by choosing and constructing an applicable fitness function.
