Comparative study of heuristics for optimal printed circuit board assembly

This paper is concerned with the machine construction of printed circuit boards (PCBs). The critical steps in this process include path planning, drilling, and part placement or insertion. To enhance production speed it is important that an optimal (or near-optimal) assembly sequence can be found which performs these tasks in a minimal time. The general PCB assembly problem is at least as complex as the traveling salesperson problem (TSP) which is known to be NP-complete, so that an exact solution using optimization theory is out of the question for any but the smallest problems. Thus a heuristic line of attack must be used, which finds a near-optimal solution in an acceptable time. The present paper concentrates mainly on three such methods: Genetic Algorithms (GA), Evolutionary Programming (EP), and Simulated Annealing (SA). A critical discussion of each technique is given and their performance on realistic PCB assembly problems is analyzed.