˺ An Experimental Approach For Optimization of Surface Quality Parameters M.A.Shouman* A.A Aboul-Nour** E.A.Elsayed** *Professor of O.R., Faculty of Computers and informatics, Zagazig University, ProfShouman@hotmail.com ** Industrial Engineering Dept, Faculty of Engineering, Zagazig University ABSTRACT New mathematical models have been developed for surface roughness errors prediction based on different parameters. Six of them are vertical parameters while the other is horizontal one. The models have built at different machining conditions. Validation analysis have been made for the empirical formulae through correlation coefficient R, determination factor R, 2 and root mean square errors RMSE to evaluate the strength between response variables and experimental measured values. The results show that the empirical formula provide nearly exact values of surface roughness error for the different response variables and are superior to be as predictable tool in terms of accuracy and engineering quality control. The validation has been made along 300 experiments. Comparison between artificial neural network ANN and mathematical model has been established as prediction technique. The results are promising. A program has been developed for predicting and optimizing geometric errors of different materials based on the data of experimental works and ANN. General remarks, tendencies, and conclusion have been presented. Keywords: Design of experiment, Form errors, Quality control, Surface roughness error. INTRODUCTION In recent years the increasing demand of high quality manufacturing parameters, (in terms of dimension, tolerances, process monitoring and control) has lead to sever need of prediction techniques for production environment in the aim of avoiding undesirable accuracy results. Consequently, the accuracy requirements for machine parts have continuously increased and tend to be specially critical in modern industry [1]. Recently the techniques of form errors for surface topography have gained importance from the point of view of their function to be performed in an assembly. The surface roughness, waviness, lay, and flaws are classified as micro geometric errors and are considered as main items for any surface quality and nominal surface is the intended surface. The nominal surface does not include intended surface roughness and a real surface is the actual boundary of an object