This study focuses on PCB scheduling problem with sequence-dependent setup times on a single machine and feeder capacity constraint. The objective of the scheduling problem is to minimize the total weighted tardiness. The evolutionary algorithms to solve this PCB scheduling problem is an integration of genetic algorithm, local search procedure, heuristic scheduling rule and Keep Tool Needed Soonest (KTNS) policy. Heuristic scheduling rules and genetic operators are applied to create chromosomes in populations. Local search procedure and the KTNS policy are employed to reduce feeder setup times. Extensive numerical experiments were conducted to compare the performance of the evolutionary algorithms—combinations of memetic or genetic algorithms and heuristics scheduling rules. The result shows that memetic algorithm with Minimum Slack Time (MST) rule in the initial population, yields better solution. High competition in electronic industry during the last decade has forced electronic product types are various; demand varies from a few parts to many thousand parts with differing part specifications. As a result of the situation, capability of processing in small production lot size is needed. Production in small lot size implies that an SMT machine must be stopped while changing feeders on the machine. This results in high setup times.
[...] The result derived from performing the procedure for the eight PCB lots is shown in Table 4. Table 3 Example of KTNS policy Component types PCB lots # of required comp onents Table 4 Example of KTNS policy PCB Lots Component types : component kept from KTNS policy : one setup component required # of required components # of component setups In order to improve quality of solutions, a local search procedure is employed to reduce the number of feeder setups as long as the total weighted tardiness does not increase. [...]
[...] Consequently, minimizing the feeder setup time is a critical issue for PCB scheduling PCB scheduling problem involves constraint of component feeder slot capacity. The total number of required feeder slots for all PCBs during a planning horizon is generally greater than the feeder slot capacity, but the number of feeder slots for a PCB lot is less than the feeder slot capacity. Thus, efficient operation management policy should be employed to reduce the number of feeder slot setups. Some previous studies considered limitation of the feeder slot capacity. [...]
[...] Step Apply the local search procedure and the KTNS policy in the new individuals and return to step Implementation of Evolutionary Algorithms In this section, the manner in which each component of the evolutionary algorithms was implemented to solve PCB scheduling problem, and input data will be discussed. The input data of each evolutionary algorithm can be categorized into four parts. Production data consists of the number of PCB types (PCB lots), component types required by each PCB lot, quantity of each component type per board, PCB lot sizes, due date of each PCB lot, and penalty cost of late delivery for each lot. [...]
[...] KTNS policy and Local Search Procedure The KTNS policy is applied to reduce component feeder setup time for a specific sequence of PCB lots. Assuming each PCB lot does not require the number of feeder slots more than feeder capacity of SMT machine. The KTNS policy and their variables are presented as follows: C O M K m r f ψ ( set of component types required by the current PCB lot set of component types that are on the machine set of component types that have to be installed on the machine set of components kept feeder capacity number of feeder slots required by the current PCB lot number of free slots number of feeder slots required by component types in set i Step Read component types required by the current PCB ( C If the current PCB lot is the first lot in the sequence, then set C = O = M and K = φ ; otherwise, determine members of set M where M = C O . [...]
[...] Knuutila, and O. S. Nevalainen; The general two-level storage management problem: A reconsideration of the KTNS-rule. European Journal of Operational Research; Vol 171(1):pp 189- T. C. Harrington, D. M. Lambert, and M. christopher; A methodology for measuring vendor performance. Journal of Business Logistics; Vol 12(1):pp 83- D. R. Sule; A heuristic procedure for component scheduling in printed circuit pack sequencers. International Journal of Production Research; Vol 30(5):pp 1191- S. M. Lee and A. A. Asllani; Job scheduling with dual criteria and [...]
APA Style reference
For your bibliographyOnline reading
with our online readerContent validated
by our reading committee
