Finding order in the design landscape of simple optical systems

Contrary to the frequent tacit assumption that the local minima of a merit function are points scattered more or less randomly over the design landscape, we have found that, at least for simple imaging systems (doublets with three and triplets with five variables) all design shapes we have observed thus far form a strictly ordered set of points, the "fundamental network". The design shapes obtained for practical specifications with global optimization algorithms are a subset of the set of local minima in the fundamental network and are organized in a way that can be understood on the basis of the fundamental network.

[1]  John R. Koza,et al.  Automated synthesis of both the topology and numerical parameters for seven patented optical lens systems using genetic programming , 2005, SPIE Optics + Photonics.

[2]  Gregory W. Forbes,et al.  Towards global optimization with adaptive simulated annealing , 1991, Other Conferences.

[3]  Florian Bociort,et al.  Saddle-point construction in the design of lithographic objectives, part 1: method , 2008 .

[4]  M. Isshiki,et al.  Lens Design: Global Optimization with Escape Function , 1995 .

[5]  Florian Bociort,et al.  Networks of local minima in optical system optimization. , 2004, Optics letters.

[6]  David Shafer,et al.  Global Optimization in Optical Design , 1994 .

[7]  Florian Bociort,et al.  Finding new local minima in lens design landscapes by constructing saddle points , 2009 .

[8]  Kendra E. Moore Algorithm for global optimization of optical systems based on genetic competition , 1999, Optics + Photonics.

[9]  Florian Bociort,et al.  Saddle-point construction in the design of lithographic objectives, part 2: application , 2008 .

[10]  Thomas G. Kuper,et al.  Global optimization for lens design: an emerging technology , 1993, Optical Systems Design.

[11]  Florian Bociort,et al.  Looking for order in the optical design landscape , 2006, SPIE Optics + Photonics.

[12]  Florian Bociort,et al.  Topography of the merit function landscape in optical system design , 2004, SPIE Optical Systems Design.

[13]  Florian Bociort,et al.  Network search method in the design of extreme ultraviolet lithographic objectives. , 2007, Applied optics.

[14]  F. Bociort Saddle points reveal essential properties of the merit-function landscape , 2008 .