Multi-objective optimization of the turning process using gravitational search algorithm (GSA) and NSGA-II approach

Abstract

In this paper we proposed a gravitational search algorithm (GSA) and a NSGA-II approach for multi-objective optimization of the CNC turning process. The GSA is a swarm intelligence method exploited the Newtonian laws on elemen-tary mass objects interaction in the search space. The NSGA-II is the evolu-tionary algorithm based on a non-dominated sorting. On the basis of varying values of the three independent input machining parameters (i.e., cutting speed, depth of cut, and feed rate), the values of the three dependent output variables were measured (i.e., surface roughness, cutting forces, and tool life). The obtained data set was further divided into two subsets for the training data, and the testing data. In the first step of the proposed approach, the GSA and the training data set were applied to modelling the suitable model for each output variable. Then the accuracies of the models were checked by the testing data set. In the second step, the obtained models were used as the objective functions for a multi-objective optimization of the turning process by the NSGA-II. The optimization constraints relating to intervals of depend-ent and independent variable were set on the theoretical calculations and confirmed with experimental measurements. The goal of the multi-objective optimization was to achieve optimal surface roughness, cutting forces, and increasing of the tool life, which reduces production costs. The research has shown that the proposed integrated GSA and NSGA-II approach can be suc-cessfully implemented not only for modelling and optimization of the CNC turning process but also for many other manufacturing processes.