Multi Model Multiple Line Balancing Problem Modeling and Real-Life Application

Abstract

This study deals with multi-line balancing for multiple products. This problem originates from a real-life problem at a global heat boiler manufacturing company. The objective is to develop a solution method that enables create efficient line balances in a dynamic environment which maximizes the weighted average line efficiency and reduces waste of time for frequent line balancing procedures. Since weighted average efficiency is a key performance indicator for the company this paper focuses on the optimization of this measure. The proposed method is demonstrated with an implementation at three assembly lines where great majority of the company’s production takes place. For this problem an algorithm has been developed that creates a set of feasible scenarios (a scenario pool) in which total demand of each model is divided between available lines under capacity restrictions. For every scenario in the pool a custom multi model assembly line balancing MIP is created automatically and run. From the results best (most efficient) assignment of models’ demand to lines and corresponding line balances are obtained. Furthermore a user-friendly dynamic decision support system (DSS) is developed and proposed solution approaches are embedded in this DSS for the daily operations of the company. The developed DSS enables users to do an efficient line balancing in short computational time and provides the results with detailed line balancing reports. As this problem can be faced in various areas the proposed solution approaches can also be applied to different sectors and factories with little modification. © 2020 Elsevier B.V. All rights reserved.