HOURS-PER-VEHICLE PRODUCTIVITY MODEL DEVELOPMENT IN AUTOMOTIVE INDUSTRY

• Main Author: Andang Tjahjono, Warih
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/52969
• Institution: Institut Teknologi Bandung

To become a major player in the global economy, automotive industry in Indonesia must have the best competitiveness; one of the most important pillars is a high productivity. An automotive industry is an industry that utilizes high technology, that in its operations involves a lot of human resources, and then high productivity is the main challenge. To cope with this challenge, an important strategy for PT. X, in the 2017-2019 period, has been planned in order to achieve a significant increase in its productivity level to be the best competitiveness, with relatively affordable investment. Intensive implementation of karakuri activities and idea suggestions is seriously considered as an important factor. Karakuri is a mechanical assistive device that can improve the quality of work (e.g. simplify, lighten, bring closer) by using simple mechanics (e.g. gravity, potential energy, etc.) without the help of electric, hydraulic, or pneumatic power and without computerization, which thus requires small investment. This study is aimed at modelling the effect of karakuri activity and idea suggestions for increasing productivity at PT. X. More specifically, the objective of this study is to identify the significant factors that influence productivity at PT. X. This study analyzes field data taken in the period July 2017 - December 2019, using hour-pervehicle (HPV) model with the Multiple Linear Regression (MLR) method and Bayesian Additive Regression Trees (BART) method. From the data analysis, it was found that MLR could not be used, because of the multicollinearity between the independent variables. Modelling is then carried out using BART, which has a machine learning algorithm. From the BART model, with the number of trees 10, the number of MCMC samples 10, and the quantile of prior 0.99, is the best model in representing the productivity of PT. X. Furthermore, several significant factors are identified such as idea suggestions, number of types, number of working hours, and karakuri to be factors as having significant influence the productivity. Therefore, through this model, we could conclude that karakuri activity and idea suggestion are significant in increasing productivity at PT. X, which then it can be used for automotive manufacturing companies in general.