Automated pre- and post processing of optical simulations using the Python scripting language

Optical simulation tasks such as illumination system design or straylight analysis are normally performed using commercial raytracers. While the latter are excellent for predicting the performance of optical systems, they are often pushed to their limits, if huge amounts of input and output data need to be handled, if system geometry variations have to be computed or if the data analysis is non-standard. Splitting the task into the input/output data management and the optical simulation itself can help. We implement the first using the standard scripting language Python as a "server". The latter is performed by the raytracer, which acts as the "client" and appears to the server only as a black box. Because the raytracer is reduced to its basic functionality, the different software products become more or less interchangeable. Clever interfaces on the server-side solve many optical problems almost independently of the underlying client. In the talk, an example is shown where we perform an optimization of an illumination system using ASAP (Breault Research Organization) as the client.