Application of particle swarm optimization to the automatic design of optical systems

Various global optimization methods are available for the automatic design of optical systems. However, these methods are not specifically tailored at freeform systems with a large number of variables. This paper presents a novel method for the automatic design of optical systems which is based on Particle Swarm Optimization (PSO). PSO was originally introduced in 1995 to model the interaction of individuals in a swarm or a flock of birds. The optimization of a problem is performed by iteratively improving a candidate solution with regard to a specific merit function. A collection of candidate solutions, called particles, move around in the search space according to simple mathematical rules. The movement of each particle is influenced by its local best known position and by the best known positions found by other particles. Repeating this process is expected to guide the swarm to the best solutions. Important aspects of PSO are the communication of the particles with each other and the ability to learn from the experience of the swarm. A PSO algorithm has been implemented in a custom optical design software. In the application to optical design, each particle represents an optical system in the multi-dimensional parameter space. The merit function is a measure for the quality of the optical system. The application of PSO is demonstrated through several examples with and without freeform elements. The results prove that the proposed method is an excellent tool for the optimization of freeform systems with a large number of variables.