Multi objective optimizing of machining processes is used to simultaneously achieve several goals such as increased product quality, reduced production time and improved production efficiency. This article presents an approach that combines grey relational analysis and regression modeling to convert the values of multi responses obtained from Taguchi method design of experiments into a multi objective model. The proposed approach is implemented on turning process of St 50.2 Steel. After model development, Analysis of Variance (ANOVA) is performed to determine the adequacy of the proposed model. The developed multi objective model is then optimized by simulated annealing algorithm (SA) in order to determine the best set of parameter values. This study illustrates that regression analysis can be used for high precision modeling and estimation of process variables.

Abstract. Gas Metal Arc Welding (GMAW) is one of the most extensively used processes in automated welding due to its high productivity. However, to simultaneously achieve several con icting objectives such as reducing production time, increasing product quality, full penetration, proper joint edge geometry, and optimal selection of process parameters, a multi-criteria optimization procedure must be used. The aim of this research is to develop a multi-criteria modeling and optimization procedure for GMAW process. To simultaneously predict Weld Bead Geometry (WBG) characteristics and Heat-A ected Zone (HAZ), a Back Propagation Neural Network (BPNN) has been proposed. The experimentally derived data sets are used in training and testing of the network. Results demonstrate that the nely tuned BPNN model can closely simulate actual GMAW process with less than 1% error. Next, to simultaneously optimize process characteristics, the BPNN model is inserted into a Particle Swarm Optimization (PSO) algorithm. The proposed technique determines a set of values for parameters and the workpiece groove angle in such a way that a pre-speci ed WBG is achieved while the HAZ of the weld joint is minimized. Optimal results are veri ed through additional experiments.