WORLD SCIENTIFIC eBooks
Information Theoretical Approaches to Quantum Chemistry
Shridhar Ramachandra Gadre
Abstract
The development of the concept of entropy is traced briefly from the time of Boltzmann to the present. The axiomatic formulation of the Shannon information entropy functional [C. E. Shannon, Bell System Tech. J. 27 , 379 (1948)], viz. S = - k Σ p i ln p i is presented. The maximum entropy principle (MEP) due to Jaynes [E.T. Jaynes, Phys. Rev. 106 , 620 (1957)] is summarized and the existence and uniqueness of such maximum entropy solution with the constraints of moments of the distribution is discussed. This is followed by an account of the practical reconstruction of the electron momentum distributions and Compton profiles by employing the first and second moment constraints. Connections of information entropy encompassing kinetic energy functional, reaction dynamics, loges as well as surprisal analysis are brought out. The pioneering contribution of Białynicki-Birula and Mycielski [I. Białynicki-Birula, and J. Mycielski, Commun. Math. Phys. 44 , 129 (1975)] highlighting the connection between the sum of information entropies in conjugate spaces and the Heisenberg uncertainty relation is outlined. The problem of characterizing the quality of atomic and molecular wave functions has been considered in terms of the sum of information entropies in conjugate spaces. This is followed by the discussion of a new maximum entropy principle [S. R. Gadre and R. D. Bendale, Int. J. Quantum Chem. 28 , 311 (1985)] stipulating the maximization of the sum of the Shannon entropies in conjugate spaces. Some applications of this new maximum entropy principle for obtaining the probability distributions in conjugate spaces are summarized.