Novel heuristic approach to integrating task sequencing and production system configuration

Abstract

The robot system integration (SI) business targets the construction of production lines for various products that differ from customer to customer. Improving the efficiency of line design is important for increasing profits in the SI business. Especially in the construction of assembly lines, the optimal system configuration, with an appropriate combination of robotic and human stations, is critical for maximizing the return on investment of customers. Yet, the optimal system configuration depends heavily on the task sequence in the process plan. In this paper, task sequencing and system configuration are regarded as two interdependent sub-problems of line design. The exact joint optimum of the two problems can be found only if they are solved together, in an integrated way. However, this is computationally intractable for problems of industrially relevant sizes. To overcome this challenge, a heuristic method exploiting industrial knowledge is proposed for solving the two problems simultaneously. The heuristic constructs promising task sequences according to the capabilities of the applicable resources, considering aspects such as, among others, human work quality or robot tool change times. Then, a close-to-optimal system configuration is computed by considering the promising task sequences and their various relevant combinations. The proposed method is illustrated in a case study on the assembly of electrical components in the automotive industry.