Optimization of robotic task sequencing problems by using inheritance-based PSO

摘要

For the optimal robot-arm multi-point manufacturing sequence (RAMMS) problem, the best sequence needs to be explored to efficiently visit each task point just once and return back to the start one finally. It is the known travel salesman problem (TSP). Furthermore, in each task point, the optimal joint configuration is also a difficult problem because it involves an inverse kinematics (IK) computation and may contain multiple configurations in the solution space. To conquer the difficult problem, this paper developed an inherence-based time-vary acceleration coefficient (IHTVAC) particle swarm optimization (PSO) and designed a two-phase evolutionary approach for the solution explorer. In the first phase, the joint configurations of robots in task points are evolutionary explored. The second phase then includes both task sequences and joint configurations into discrete-type evolutionary loop to explore the global solutions. The experimental result shows the advantage of the proposed two-phase approach.