Interval Algorithm for the Global Search of Uncertainty Domains in Simulating the Kinetics of Complex Chemical Reactions

This study is devoted to the mathematical description of a kinetic experiment under the setting of multiple solutions to the inverse problems of chemical kinetics. A new algorithm for the global search of uncertainty domains is proposed, which represents a symbiosis of the methods of global optimization with an interval evolutionary approach. Chemical reaction models built using the kinetic parameters of these areas describe experimental data within an acceptable measurement error specified by the researcher. The use of this algorithm allows a complex multidimensional domain in the global space of kinetic parameters, whereas the local solution sets describing the experimental data with near equal accuracy are searched or the one-, two-, or three-dimensional projections of uncertainty domains are separately built in many other studies. The proposed algorithm has been tested using a reaction simulation of n-decane inhibited oxidation. Since this structure has a formal parallelism, parallel computing techniques are applied to speed up calculations.