Comment concerning the optimum control of transformations in an unbounded quantum system

Abstract

We discuss the achievability of optimal control of the evolution of an unbounded quantum system by the action of external fields. Previous work has established that the evolution of a quantum system with a nondegenerate discrete and bounded spectrum of states can, in principle, be fully controlled, i.e., that the system wave function can be guided by external fields to approach arbitrarily closely a selected target state wave function. The optimal control of the evolution of a quantum system with a discrete and bounded spectrum of states has also been studied, using a method of analysis that depends on the localized character of bound state wave functions and the fact that the spectrum of states is bounded. In this paper, we examine whether it is possible to control, partially or fully, the evolution of an unbounded quantum system. We show that optimal control of the evolution of an unbounded quantum system is possible, in the sense that it is possible to minimize the difference between the product function formed at time t0 from a localizing function and a continuum wave function and a similarly defined target function at time tf. We have not been able to establish that such optimal control is equivalent to full control, i.e., that the difference between the initial and target functions can be made arbitrarily small, which would require showing that the set of control functions is complete with respect to the function space of the system. Our analysis establishes the existence of external fields that can optimally guide the unbounded system evolution in the absence of other constraints than the Schrödinger equation of motion, but does not provide an algorithm for the construction of such fields. The relationship between optimal control of the evolution of a quantum system and the existence of chaotic dynamics in the continuum domain is briefly discussed. We argue that the optimally controlled system cannot have chaotic dynamics even if the bare system does. As illustrations of the results obtained, we discuss briefly the optimal control of evolution in the subset of discrete states from a spectrum with both discrete and continuum states, the optimal control of evolution in resonant states, and the inversion of the optimally controlled localized product continuum target function to yield the corresponding system wave function.