What is an ill-posed inverse problem? The answer to this question is the main objective of the present paper and the pre-requisite to follow the material requires only elementary calculus. The first mathematical formulation of an inverse problem, due to N. H. Abel, together with the fundamental work by Jacques Hadamard, are explored at the beginning of the paper. A prototype system is used to consider the regularization concept. Three numerical methods, the Tikhonov regularization, the decomposition into singular values and the Hopfield neural networks, applied to remove the singularity are examined. General aspects of the ill-posed inverse problems in chemistry with emphasis in thermodynamics and a set of general rules for other areas of science are also analyzed.

Um algoritmo geral para resolver problemas inversos lineares e não lineares, baseado em redes neurais recursivas, é discutido neste trabalho. O procedimento será aplicado a problemas físico-químicos modelados por equações integrais, diferenciais e de autovalor. As aplicações são discutidas em espectroscopia de aniquilação de pósitrons, cinética química e espectroscopia vibracional. O método é robusto com relação a erros nas condições iniciais ou em dados experimentais. A presente abordagem é simples, numericamente estável e tem uma grande aplicabilidade.

A general algorithm to solve linear and nonlinear inverse problems, based on recursive neural networks, is discussed in this work. The procedure will be applied to physical chemical problems modeled by integral, differential and eigenvalue equations. Representative applications discussed are in positron lifetime spectroscopy, chemical kinetics and vibrational spectroscopy. The method is robust with respect to errors in the initial condition or in the experimental data. The present approach is simple, numerically stable and has a broad range of applicability.

A simple analysis of glory and rainbow effects, together with the description of their trajectories is given by the geometric model. The energy dependence of glory and rainbow impact parameters and the energy dependence of the rainbow angle are determined analytically within the model. An universal function for glory and rainbow trajectories can be easily determined.