Ryo Murakami, Sachio Kobayashi, Hiroki Kobayashi and Junji Tomita
Automating electronics assembly is complex because many devices are not manufactured on a scale that justifies the cost of setting up robotic systems, which need frequent readjustments as models change. Moreover, robots are only appropriate for a limited part of assembly because small, intricate devices are particularly difficult for them to assemble. Therefore, assembly line designers must minimize operational and readjustment costs by determining the optimal assignment of tasks and resources for workstations. Several research studies address task assignment issues, most of them dealing with robot costs as fixed amount, ignoring operational costs.
In real factories, the cost of human resources is constant, whereas robot costs increase with uptime. Thus, human workload must be as large and robot workload as small as possible for the given number of humans and robots. We propose a new task assignment method that establishes a workload balancing that meet precedence and further constraints. The following must be determined before using our method: which tasks robots can perform, and which workstations robots are assigned to. We assume that humans can perform every task and consider the constraints that restrict the tasks robots can perform. By applying our method to several case studies, problems involving 20 humans were solved within 1 minute and 1% dispersion.
These results indicate that our method can be used in actual factories where a short-term planning period corresponding to frequent production fluctuations is required. We also applied our method to real assembly data for laptops manufactured by our company and obtained task assignment that reduces the operational costs by 30%. This suggests that our method can contribute to promoting the automation of electronics assembly by demonstrating its cost reduction potential.
In this paper, we presented a new method of handling an assignment problem for hybrid manual and robotic assembly mixing lines. The method is based on a tabu search for a constraint satisfaction problem. The aim is to assign tasks to workstations, each of which is occupied by either a human or a robot. The focus is on how to deal with the idle time of each workstation. By assigning tasks to the workstations of humans so that no idle time remains, our method reduces the total processing time of robots by approximately 40% in the laptop case study presented. In the future, further research will be undertaken on reassigning existing assembly lines in response to changes to the production process or production program.
Initially Published in the IPC Proceedings