Similarity Analysis For Assembly Line Balancing

Most of the traditional assembly line balancing algorithms are performing the clustering of operations on the basis of estimated operating times with the objective of minimizing the workstation idle time. Such an approach gives relatively good results for manual assembly lines but is less pertinent for automated and/or heterogeneous assembly systems.
This paper presents a new approach for clustering assembly operations that could be used at an early assembly Une design stage. The proposed procedure exploits the introduction of Working Principles (WPs) and performs the clustering of these WPs on the basis of an affinity coefficient. A new clustering algorithm has been developed which allows to introduce assembly workstation design concerns in the proposed approach.
This original approach for assembly line design is described and illustrated on a case study.