An evolutionary simulated annealing algorithm for optimizing robotic task point ordering

High productivity requires that robot manipulators perform complex tasks in minimum time. The paper presents an evolutionary simulated annealing (ESA) algorithm for optimizing an important class of complex tasks of point-to-point moves, such as mechanical assembly, electronic component insertion, and spot welding. This algorithm combines the basic principles of two major heuristic search methods: simulated annealing and genetic algorithms. Indeed, these methods are commonly used to solve the well-known traveling salesman problem (TSP), and the point ordering problem in robotics is very similar to the TSP in mathematics. The three algorithms have been implemented in a computer-aided design (CAD) software system, CATIA. Experimental results show success factors for using ESA.