January 28, 2009
3 pm in Howey Physics Room N110

 

Raj Chakrabarti
Department of Chemistry, Princeton University

"Control Landscapes in Biology and Physics"

 

A control landscape is the map between the control variables and a target objective function. This talk will review our studies on control landscapes in I. Evolutionary Biology and II. Quantum Physics.

I. Evolutionary Biology. The focus here is on fitness landscapes where protein sequences serve as the controls. The evolutionary dynamics of proteins may be understood as adaptive walks on such fitness landscapes. Work will be presented  on quantitatively identifying the fitness measures that guide the evolution of natural enzymes. Computational studies will be presented providing an understanding of the observed sequence distributions of natural enzyme families based on these measures.

II. Quantum Physics.  Quantum mechanical landscapes will be considered for the coherent manipulation of molecular dynamics using laser fields as controls. Extensive laboratory evidence indicates that optimal quantum controls are surprisingly easy to find. The landscape for coherent control of population transfer is devoid of local traps given a suitably diverse set of controls.  This finding rationalizes the prior laboratory successes and lays the foundation for the application of coherent control to complex problems, including quantum computation realizations.

The relationship between these diverse topics lies in their common expression in terms of control landscapes.  Further potential implications of such landscapes will be discussed.