Optimization of a Real Time Web Enabled Mixed Model Stochastic Assembly Line to Reduce Production Time

Abstract

The role of assembly lines has never been more critical as it is now with the world entering the 4th Industrial Revolution, commonly referred to as Industry 4.0. If the focus of the previous industrial revolution was on mass production, the focus of Industry 4.0 is on mass customization. One of the major changes mass customization brings about to an assembly line is the need for them to be autonomous. An autonomous assembly line needs to have the following key features; ability to provide a ubiquitous input, the ability to optimize the model in real time and achieve product variety. Product variety, in this context, refers to different variants of the same product as determined by the user. Assembly lines that make provision for introducing product variety are termed as mixed-model assembly lines. Mixed-model assembly lines become stochastic in nature when the inputs are customized as time cannot be predetermined in a stochastic process. The challenge, as it stands, is that there are limited discussions on real-time optimization of mixed model stochastic assembly lines. This paper aims to highlight this challenge by considering the case study of optimizing a mixed model assembly line in the form of a water bottling plant. The water bottling plant, which needs to produce two variants of the bottled water, 500 ml, and 750 ml, takes customer inputs through a web interface linked to the model, thereby making it stochastic in nature. The paper initially details how the model replicating the functioning of the water bottling plant was developed in MATLAB. Then, it proceeds to show how the model was optimized in real time with respect to certain constraints. The key results of the study, among others, showcase how the optimization of the model is able to significantly reduce production time.