Abstract:
The uniqueness and stability of the solutions of two- and three-parameter inverse problems in nonlinear fluorimetry (saturation fluorimetry) of complex organic compounds is studied by the method of computer simulations. The main attention is paid to a practical stability of such problems with respect to the input data noise and variations in the model. The use of the technique of artificial neural networks (including their training on noisy input data) provides the solution stability under real experimental conditions.