Redundant configuration of robotic assembly lines with stochastic failures

Abstract

One of the main challenges in the operation of robotic assembly lines is the occurrence of failures. Due to the connection of the stations via a material handling system, failures at one station often result in throughput losses. To some extent, these throughput losses can be reduced by installing buffers between the stations. However, the installation of buffers requires considerable investments and scarce factory space. Due to the advances of manufacturing technologies that form the foundation of ‘Industry 4.0’, new solutions to reduce failure-related throughput losses open up. One solution is a redundant configuration, in which downstream (backup) stations automatically take over the operations of failed stations during repair time. The throughput loss in these situations depends on the allocation of operations and the assignment of backup stations. Existing approaches in the literature that consider redundancies in the configuration of automated lines neglect the resulting production rate. Instead, the lines’ level of redundancy is used as a surrogate measure for optimisation. We present a genetic algorithm for the redundant configuration of robotic assembly lines with stochastic failures to maximise the production rate of the line. In a numerical analysis, it is demonstrated that this approach allows for productivity improvements.