Using a random restart hill-climbing algorithm to reduce component assembly time in printed circuit boards

Abstract

This paper presents a heuristic for reducing component placement time in assembling a printed circuit board using the random restart hill-climbing method. The assembly problem is divided into four different and related sub-problems: definition of the component placement sequence, definition of the distribution of the components in the magazines, definition of the pickup sequence, definition of the nozzle sequence. The random restart hill-climbing method was used in two different stages: in the first stage to optimize the component placement sequence and the component distribution sequence in the magazines, in a global optimization process; in the second stage, to optimize the placement sequence and the pickup sequence in the local process. The results show the best performance of the proposed method compared with three other methods: one using a genetic algorithm, another using a memetic algorithm and a third solution provided by machine manufacturing.