• Corpus ID: 14149737

Optimization of Optical Structures Using Markov Decision Processes

  title={Optimization of Optical Structures Using Markov Decision Processes},
  author={Tyler W. Hughes and Yu Shi},
We present an algorithm for optimizing optical structures based on Markov Decision Processes. Our method is successfully demonstrated on a one-dimensional stack of dielectric slabs, where we wish to choose a combination of slab thicknesses and refractive indices to approximate a target reflection spectrum. We show that this algorithm is less susceptible to being caught in local minimum and has favorable computational scaling as the number of layers is increased. 

Figures from this paper



Method for sensitivity analysis of photonic crystal devices.

We present a new method for sensitivity analysis of photonic crystal devices. The algorithm is based on a finite-difference frequency-domain model and uses the adjoint variable method and

Inverse design of nanophotonic structures using complementary convex optimization

The target field, which started as a sinusoid in a Gaussian-envelope, closely matches the field obtained from simulating the dielectric structure using FDTD, and the values of Y were strictly constrained to be between 1 and 10, which resulted in a nearly binary dielectrics structure.

Double layer anti-reflective coatings for silicon solar cells

In this paper simulated single and double layer anti-reflective coatings based on the refractive index limits of silicon nitride (SiN) and silicon oxide (SiO/sub 2/) are presented. The best structure

Passive radiative cooling below ambient air temperature under direct sunlight

An integrated photonic solar reflector and thermal emitter consisting of seven layers of HfO2 and SiO2 that reflects 97 per cent of incident sunlight while emitting strongly and selectively in the atmospheric transparency window demonstrates that the cold darkness of the Universe can be used as a renewable thermodynamic resource, even during the hottest hours of the day.