Efficient optimization of illumination systems with differential raytracing in combination with particle swarms and evolution algorithms

The fast and efficient optimization is of great importance for today's optical illumination systems design. Contrary to the merit function based on the Monte-Carlo raytracing technique, a new merit function based on differential ray tracing is derived. As the differential approach works noiseless the resulting merit function excels through a smooth and continuous form. Hence, differential raytracing possesses great potential to improve the optimization efficiency by reducing computational time while concurrently showing a constant optimization quality. We review the approaches implemented in the optical design software RAYTRACE using the example of optimizing the intensity distribution in a square target plane to be homogeneously. Additionally, we present comparative calculations between deterministic (e.g. gradient methods) and stochastical optimization techniques (e.g. Monte-Carlo method, simulated annealing, evolution algorithms, particle swarm optimization) employing the merit function based on differential raytracing.