Integrated task sequencing and path planning for robotic remote laser welding

This paper investigates the problem of integrated task sequencing and path planning in Remote Laser Welding (RLW). It is shown that finding the appropriate order of welding tasks is crucial for exploiting the efficiency of this new joining technology, and this can be achieved only if the robot path is considered already at the time of sequencing. For modelling the problem, a novel extension of the well-know Travelling Salesman Problem with neighbourhoods and durative visits, denoted as TSP-ND, is introduced. Basic properties of this problem are formally proven, and a GRASP meta-heuristic algorithm is proposed for solving it. Extensive computational experiments demonstrate that the novel approach solves efficiently industrially relevant problems, and it achieves substantial improvement in cycle time compared to the single earlier approach in the literature dedicated to RLW, as well as compared to a decomposition approach to solving the TSP-ND model.