Composite Modeling for Evaluation of Human Based Inspection Systems

Abstract

Recent advancements in the field of manufacturing and service industries could not replace human involvement as every manufacturing system cannot undergo autonomation due to certain limitations such as budget, space, and skilled-labor. However, this autonomation alters the nature of human work. Industries do exist where the role of human labor is still vital in achieving product quality and system reliability. This scenario drove researchers to study human based quality systems resulting in an important constituent of modern manufacturing setups. Inspection process and quality inspectors are very much important for manuallly regulated systems generating the necessity of subtle human judgment for the skilled, semi-skilled, or low-skilled inspectors. Since job in complex manufacturing scenario is assigned based on varying skill-levels of inspectors, therefore, their competencies should be evaluated properly before assigning any task. That major research gap is bridged with this research by investigating visual inspection, and the effect of related factors on inspectors’ skill exploring the overall inspection performance. The entire study has been carried out in three distinct phases while adopting composite modeling. In the first phase, factors of visual inspection, those influence the human inspection skills, have been evaluated using statistical techniques. In this context, firstly the dominant factors of visual inspection were identified from the literature and the latent constructs are defined with their observed variables, which developed the basis of the theoretical model. Thereafter, the survey instrument was designed based on the validated content, and data were collected from the (value added) textile industry of Pakistan. A statistical technique, Confirmatory Factor Analysis (CFA), was then employed to analyze the collected data. The outcome was a first order model that revealed the interrelationships of latent constructs and their respective observed variables. Further, in the second phase, the results of CFA model were used to design a scale to classify inspectors into different skill levels. In this regard, Competency Assessment Model (CAM) was proposed to determine the competency score of inspectors based on the influencing factors from CFA model. One of Multi-Criteria Decision-Making (MCDM) techniques, Analytical Hierarchy Process (AHP) was used to model the selected factors and their respective weights (local values) were determined by pairwise comparison. The global values were obtained using these weights which were then used to determine the competency score of each inspector. The proposed model classified the inspectors into their skill level based on their competency score according to the pre-defined cut-off values. Finally in the third phase, a multi-objective optimization model was developed to determine the number of inspectors of each skill level to meet the demands of inspection station. The objective functions included inspection cost, outgoing quality, and inspected quantity. The proposed model considered inspection time along with inspection skill and determined the optimal values of decision variables for three different products using minmax variant of goal programming. Numerical examples with three different types of products, graphical illustration, and sensitivity analysis were carried out to illustrate the significance of model for human based offline inspection. Inspection station must have a suitable combination of quality inspectors having different skill levels to meet the requirements of inspection performance. The managers of different industries may derive benefit from the findings of this study because it is helpful for efficient utilization of human labor for quality inspection or quality control.