AN INTELLIGENT SCHEDULING SYSTEM FOR SMT ASSEMBLY LINES

SMT assembly lines are gradually replacing conventional printed circuit boards production lines. Manufacturing oriented to the client's specific needs has led to smaller lot sizes and higher set up costs. This article deals with the production scheduling problem on an SMT production line with one machine, four feeder tables (which allow off-line set up), two operators and n boards. The objective is to determine a sequence of boards (and corresponding configurations of feeder tables) which minimizes the total production time. The solution of this problem was divided into two steps: group formation and group scheduling. In group formation we group similar boards. This task involves the calculation of similarity coefficients between boards and a clustering procedure. In group scheduling, we determine the sequence of groups that minimizes the makespan. A simulated annealing approach was developed to solve the scheduling problem. Experimental results validated the proposed procedure.