I will present a new system helping architects in the design of interior lighting, combining Blender with an optimization technique called "genetic algorithm". A big challenge for the designer is to imagine the final effect of a proposed lighting configuration, before being physically realized.
The tool I'm going to present does much more than showing the designer the final effect of a single solution. Blender is embedded in a genetic algorithm optimizer, that evolves generations of hundreds of possible "solutions", selecting the best performing ones. A solution is a feasible configuration of lamps placed in the interior space, taking into account the constraints on where in the room it is possible to install lamps.
Following the analogy of natural evolution, solutions are "mated" to generate offsprings, and only the best fit individuals survive to the next generation. The fitness, in this case, is a set of measures of light quality and consumption, gathered from Blender rendering.