Abstract—Dimensional changes because of shrinkage is one of the most important problems in production of plastic parts using injection molding. In this study, effect of injection molding parameters on the shrinkage in polypropylene (PP) and polystyrene (PS) is investigated. The relationship between input and output of the process is studied using regression method and Analysis of Variance (ANOVA) technique. To do this, existing data is used. The selected input parameters are melting temperature, injection pressure, packing pressure and packing time. Effect of these parameters on the shrinkage of above mentioned materials is studied using mathematical modeling. For modeling the process, different types of regression equations including linear polynomial, Quadratic polynomial and logarithmic function, are used to interpolate experiment data. Next, using step backward elimination and 95% confidence level (CL), insignificant parameters are eliminated from model. To check validity of the PP model, correlation coefficient of each model is calculated and the best model is selected. The same procedure is repeated for the PS model. Finally, optimum levels of the input parameters that minimize shrinkage, for both materials are determined. Invasive Weed Optimization (IWO) algorithm is applied on the developed mathematical models. The optimization results show that the proposed models and algorithm are effective in solving the mentioned problems. Index Terms—IWO algorithm, Optimization, Plastic injection molding, Regression, shrinkage. I. INTRODUCTION Nowadays, competitive market requires producers to produce high quality parts, with lower price in the least possible time. Injection molding is known as an effective process for mass production of plastic parts with complicated forms and high dimensional precision. In this method, high pressure fluid polymer is injected to the cavity with desired form. Next, under high pressure, fluid solidifies. During the process, plastic materials are under high pressure and temperature. Materials are cooled to get desired form. Injection molding process can be divided into four stages: Plasticization, injection, packing and cooling. Although molding process may seem simple, the molded polymers are effected by many machine parameters and process condition [1-2]. Manuscript received March 31, 2011; revised May 24, 2011. Alireza Akbarzadeh is with the Department of Mechanical Engineering, Ferdowsi University of Mashhad, Mashhad, Iran.Tel/Fax: 98-511-876-3304; E-mail: Ali_Akbarzadeh_T@ yahoo.com. Mohammad Sadeghi is with the Department of Mechanical Engineering, Ferdowsi University of Mashhad, Mashhad, Iran.E-mail: H.Sadeghi@ymail.com. Incorrect input parameters settings will cause bad quality of surface roughness,decreases dimensional precision, Warpage, unacceptable wastes, increases lead time and cost [3].Therefore, finding the optimized parameters is highly desirable. In past scientists used trials and error to find good process conditions but this method is time and cost consuming. In addition, when there are a large number of input parameters, these methods can’t be used. Nowadays, the model of the process and optimal condition are developed using analytic methods and heuristic algorithms [4-8]. In previous studies, critical parameters that affect the quality of the parts are investigated. Hang et al. [4] considered six input parameters as; mold temperature, melting temperature, packing pressure, packing time and injection time. They studied effects of these parameters on surface quality of the thin molded parts. Li yang et al. [5] investigated effect of the same parameters with the addition of injection speed, injection acceleration on width of the segregation line. Chang et al. [6] studied effects of melting temperature, injection temperature, packing time and packing pressure on the surface quality of the produced parts using fuzzy logic. Sue et al. [7] used Artificial Neural Network (ANN) and SA algorithm to optimized surface quality of produced parts. Shi et al. [8] used numerical simulation and Genetic Algorithm (GA) to achieve best shear stress. Warpage in plastic parts due to anti-symmetric shrinkage is one of the most important defects caused by residual stress. These stresses are usually due to the one directional anti-symmetric shrinkage. As the shrinkage decreases, shrinkage in 3 directions decrease and therefore warpage decreases [9]. Prediction of shrinkage is very difficult because of the number of parameters and complexities of the process. Despite huge studies on modeling and optimizing of injection molding process, a few researches deal with PP and PS produced parts. Altan [10] utilized Taguchi method to optimize shrinkage of plastic, PP and PS, injection molding parts. He also applied neural network to model the process and was able to achieve 0.937% and 1.224% shrinkage in PP and PS, respectively. In this paper, we extend the Design of Experiment (DOE) study performed by Altan [10] by developing a regression model and applying IWO algorithm to obtain the optimum levels. We show that our method results in slight improvement in lowering shrinkage. This paper is organized as follows. First experimental data and selected material is introduced. Regression analysis is performed and 1 st and 2 nd orders as well as logarithmic models are developed. ANOVA is used to determine significant model parameters. Finally, IWO algorithm is used to optimize input parameters to achieve desired output, minimum shrinkage. Parameter Study in Plastic Injection Molding Process using Statistical Methods and IWO Algorithm Alireza Akbarzadeh and Mohammad Sadeghi International Journal of Modeling and Optimization, Vol. 1, No. 2, June 2011 141