Developing Statistical Models to Assess Productivity in the Automotive Manufacturing Sector

Abstract

The purpose of this study is to identify the most important activity in a value chain, effective factors, their impact, and to find estimation models of the most well-known productivity measurement, Hours per Vehicle (HPV), in the automotive industry in North American manufacturing plants. HPV is a widely recognized production performance indicator that is used by a significant percentage of worldwide automakers. During a comprehensive literature review, 13 important factors that affect HPV were defined as launching a new vehicle, ownership, car segment, model types, year, annual available working days, vehicle variety, flexibility, annual production volume, car assembly and capacity (CAC) utilization, outsourcing, platform strategy, and hourly employee's percentage. Data used in this study was from North American plants that participated in the Harbour's survey from 1999 to 2007. Data are synthesized using a uniform methodology from information supplied by the plants and supplemented with plant visits by Harbour Consulting auditors. Overall, there are 682 manufacturing plants in the statistical sample from 10 different multinational automakers. Several robust and advanced statistical methods were used to analyze the data and derive the best possible HPV regression equations. The final statistical models were validated through exhaustive cross-validation procedures. Mixed integer distributed ant colony optimization (MIDACO) algorithm, a nonlinear programming algorithm, that can robustly solve problems with critical function properties like high non-convexity, non-differentiability, flat spots, and even stochastic noise was used to achieve HPV target value. During the study period, the HPV was reduced 48 minutes on the average each year. Annual production volume, flexible manufacturing, outsourcing, and platform strategy improve HPV. However, vehicle variety, model types, available annual working days, CAC, percentage of the hourly employees, and launching a new model penalize HPV. Japanese plants are the benchmark regarding the HPV followed by joint ventures and Americans. On average, the HPV is lower for Japanese and joint ventures in comparison to American automakers by about 1.83 and 1.28 hours, respectively. Launching a new model and adding a new variety in body styles or chassis configurations raises the HPV, depending on the car class; however, manufacturing plants compensate for this issue by using platform sharing and flexible manufacturing strategies. While launching a new vehicle common platform sharing, flexible manufacturing, and more salaried employees (lower hourly) strategies will help carmakers to overcome the effect of launching new vehicles productivity penalization to some extent. The research investigates current strategies that help automakers to enhance their production performance and reduce their productivity gap. The HPV regression equations that are developed in this research may be used effectively to help carmakers to set guidelines to improve their productivity with respect to interna...