A study on the factors causing bottleneck problems in the manufacturing industry using principal component analysis

Abstract

There are a myriad of bottleneck variables that constrict the overall manufacturing capacity and make improvement decisions complex. Consequently, this study aims to identify and analyse the copious variables to pinpoint the key variable factors that influence and turn the manufacturing elements into bottleneck problem to prioritize process improvement effort. The study is limited to identifying and analysing the numerous bottleneck variables to gain insight into how much each of the variables influences the process output via the manufacturing elements. The 76-bottleneck variables abstracted from the literature were used to craft a structured questionnaire that were administered to respondents in the manufacturing industry whose size was determined at a 95% confidence level and 5% error margin respectively. The 95% confidence level is chosen to ensure adequate representation of the population and to validate the data for the study. The respondents' scores were collated into ( m x n ) data matrix which served as input variable into the factor analysis model. StatistiXL software was then employed to evaluate the data matrix. The trivial variables were discarded and 19 factors with eigenvalues (λ ˃ 1) were extracted and creatively labelled for interpretation. The result established that the “Process capability index” is the principal bottleneck factor that loaded 25% of the variables studied. The principal variables in the cluster include Equipment failure = -0.832, Operations = -0.780, Material unavailability = -0.811, and Market demand = -0.739 among others. Similarly, Manufacturing process restraint, Resources, Weather, Communication, Logistics, and Line dedication are other key factors by the magnitude of their respective variables’ factor loadings such as Random event = 0.812, Raw materials flow = -0.834, Process technology = 0.878, and Random environmental factors among other variables. Although bottleneck problems vary from one manufacturing system to another, the problems identified and the solutions presented in this study are generic and the improvement effort should focus on addressing the principal variables while not neglecting the middling and weakling variables.

Identifying Key Reliability Factors in Micro-Grid Systems Using Principal Component Analysis

“…Eigenvalues helped to retain principal component during varimax rotation in the PCA technique that was deployed for the study. The variance in the data and the decision-making process for the number of retained principal component of the reliability variable factors were measured using the eigenvalues in the PCA [38]. In the principal component analysis, the eigenvalues helped to determine only the number of principal component to be retained.…”

“…Eq. ( 5) was deployed to determine the respresentative population size of the power system population, selected for the research and that justified an adequate population size for the paper [38]. The sample size is vital and directly impacts on the accuracy to which a study generalizes findings to the larger poputaltion.…”

“…The sample size is vital and directly impacts on the accuracy to which a study generalizes findings to the larger poputaltion. Hence the sample size of considering 33 variables was taken to be large enough and adequate for the PCA study, to ensure that the sample covaraince matrix is a good approximate of the population covaraince matrix [38].…”