A hybrid approach for processing parameters optimization of Ti-22Al-25Nb alloy during hot deformation using artificial neural network and genetic algorithm

In the present investigation, isothermal compression tests of Ti-22Al-25Nb alloy were carried out under various hot deformation conditions, including the deformation temperature range of 940–1060 °C and the strain rate range of 0.01–10 s^?1. The constitutive relationship of Ti-22Al-25Nb alloy was developed using artificial neural network (ANN). During training process, standard error back-propagation algorithm was employed in the network model using experimental data sets. Based on the fitness function obtained from established ANN model, the optimization model of hot processing parameters for Ti-22Al-25Nb alloy was successfully created using genetic algorithm (GA). The optimal results achieved from the integrated ANN and GA optimization model were tested by using processing map. Consequently, it can be suggested that the combined approach of ANN and GA provides a novel way with respect to the optimization of processing parameters in the field of materials science.