This paper presents a computer-aided technology for the synthesis of molecules with optimal properties. The approach combines stochastic optimisation and group-contribution methods to search and select chemicals of desirable performance. The stochastic search is in the form of a simulated annealing algorithm. For the application of the annealing algorithm the molecular system is mapped in terms of problem states. A set of moves is applied to generate alternatives and monitor the synthesis search. Molecules are represented by molecular and composition vectors of UNIFAC groups, but most appropriate group-contribution methods are also considered. Size and composition of the molecular vector are optimised on the basis of the selected synthesis criteria without limitations on the type of calculations required. The application of the approach is illustrated with three synthesis examples. (C) 2000 Published by Elsevier Science Ltd. All rights reserved.This paper presents a computer-aided technology for the synthesis of molecules with optimal properties. The approach combines stochastic optimization and group-contribution methods to search and select chemicals of desirable performance. The stochastic search is in the form of a simulated annealing algorithm. For the application of the annealing algorithm the molecular system is mapped in terms of problem states. A set of moves is applied to generate alternatives and monitor the synthesis search. Molecules are represented by molecular and composition vectors of UNIFAC groups, but most appropriate group-contribution methods are also considered. Size and composition of the molecular vector are optimized on the basis of the selected synthesis criteria without limitations on the type of calculations required. The application of the approach is illustrated with three synthesis examples.
