“…Best conditions of cutting resulting from genetic algorithm are conformed to the measurements of experiments. A primary limitation of response surface type of approaches (whether linear such as in Lalwani, Mehta, and Jain (2008); Diniz and Micaroni (2002)) or nonlinear such as in Oktem, Erzurumlu, and Kurtaran (2005); Durairaj and Gowri (2013); Kant and Sangwan (2014) is that the obtainable results are highly dependent on the quality and accuracy of the response surface, which can be difficult to establish without a large number of experimental samples. Oftentimes however, a response surface is a "necessary evil" in order to allow use of advanced stochastic optimization approaches such as genetic algorithm, because otherwise the number of sample experiments needed becomes unreasonably large (several thousands).…”