Submission Date

7-20-2017

Document Type

Paper

Department

Physics & Astronomy

Faculty Mentor

Thomas Carroll

Comments

Presented during the 19th Annual Summer Fellows Symposium, July 21, 2017 at Ursinus College.

Supported by a National Science Foundation grant (PHY-1607335).

Project Description

A Rydberg atom is an atom with a highly excited and weakly bound valence electron. A widespread method of studying quantum mechanics with Rydberg atoms is to ionize the electron and measure its arrival time. We use a Genetic Algorithm (GA) to control the electron's path to ionization. The Rydberg electron's energy levels are strongly shifted by the presence of an electric field. The energy levels shift and curve, but never cross. At an avoided crossing the electron can jump from one level to the next. By engineering the electric field's time dependence, we thereby control the path to ionization.

A GA is an optimization method modeled on natural selection. We use a GA to evolve electric field pulses to achieve a target path to ionization. Our algorithm initially generates random members of a population and then assigns each member a fitness score based on how well they achieve our target solution. We use elitism to pass the best members of the population directly into the next generation. We use tournament style selection to choose fit members to mate and pass their genes to the next generation. We then mutate the offspring to provide genetic diversity to our population. We present our results on the effects of varying GA parameters and modifying the GA to better model the experiment.

Share

COinS