In this paper I describe a new and exciting application of optimization technology. The problem is to design a space telescope capable of imaging Earth-like planets around nearby stars. Because of limitations inherent in the wave nature of light, the design problem is one of diffraction control so as to provide the extremely high contrast needed to image a faint planet positioned very close to its much brighter star. I will describe the mathematics behind the diffraction control problem and explain how modern optimization tools were able to provide unexpected solutions that actually changed NASA's approach to this problem. Date: February 3, 2006. Key words and phrases. linear programming, convex optimization, nonlinear programming, Fourier transform, high-contrast imaging, Terrestrial Planet Finder. The author was supported by a grant from the ONR (N00014-98-1-0036). 1
Robert J. Vanderbei