| This research investigates operational, or short-term, planning for electronic assembly. A medium-demand, medium-variety printed circuit board (PCB) population will be produced on a mixed-model assembly line with two identical machines. The objective is to maximize throughput, which is the number of PCB types produced daily. The operational planning problem is decomposed into three smaller problems that are solved hierarchically. Given weekly PCB requirements, the operational planning problem determines the subset of PCBs to produce daily (part selection problem), sequences the PCBs (sequencing problem), and assigns each PCB type's components to the machines (workload balancing problem). The problems are modeled as mixed integer linear programming problems.;Heuristic procedures are developed for the part selection, sequencing and workload balancing problems. For the part selection problem, a clustering-based heuristic that incorporates similarity of both PCB lot due dates and component requirements is developed. For the daily sequencing problem, the subset of the PCBs assigned daily (by part selection problem) is sequenced based on the due date (time of day) of the lot of PCBs. For workload balancing, a heuristic is developed to bring together a workload balancing heuristic such that the solution of the workload balancing problem is optimal with respect to Johnson's two-machine scheduling problem. Numerical results are summarized for randomly generated data, and the heuristics efficiently solved various problem sizes. |
