Rancangan penjadwalan job shop pada produk stamping parts dengan metode backpropagation neural network (BPNN) untuk minimasi mean tardiness pada PT.XYZ
P T.XYZ merupakan perusahaan yang memproduksi komponen otomotifseperti dies, stamping parts, dan komponen otomotif lainnya. Perusahaanmengalami keterlambatan (tardy) dalam menyelesaikan job sehingga penyelesaianjob melewati due date. Penelitian dilakukan pada bulan Februari 2019, saat ituterdapat 11 job yang memiliki routing yang berbeda-beda dan diproduksimenggunakan 4 mesin press. Tujuan dilakukan penelitian ini adalah merancangpenjadwalan job shop pada produk stamping parts dengan metode algoritmapenjadwalan aktif, algoritma non delay, dan Back Propagation Neural Network(BPNN) dengan greedy job alignment procedure berdasarkan minimasi meantardiness. Perhitungan BPNN didukung dengan aplikasi bahasa pemrograman C#.Hasil dari perancangan penjadwalan job shop dengan ketiga metode tersebut, akandilihat metode mana yang menghasilkan minimasi mean tardiness terbaik. Hasilpenjadwalan job shop perusahaan pada produk stamping parts, terdapat meantardiness sebesar 61691 detik. Berdasarkan perhitungan yang telah dilakukan,algoritma penjadwalan aktif meminimasi mean tardiness sebesar 13,69% yaitu8443 detik, algoritma non delay meminimasi mean tardiness sebesar 27,38% yaitu16894 detik, dan BPNN dengan greedy job alignment procedure meminimasi meantardiness sebesar 49,58% yaitu 30587 detik. Hasil yang diperoleh menunjukkanbahwa minimasi mean tardiness terbaik terdapat pada rancangan BPNN dengangreedy job alignment procedure.
P T. XYZ is a company that produces automotive components such as dies,stamping parts, and other automotive components. The company experienced adelay (tardy) in completing the job so that the completion of the job beyond duedate. The research was conducted in February 2019, when there were 11 jobs thathave different routing and were produced using 4 press machines. The purpose ofthis research is to design job shop scheduling on stamping parts products withactive scheduling algorithm, non delay algorithm, and Back Propagation NeuralNetwork (BPNN) with greedy job alignment procedure based on the minimizationof mean tardiness. BPNN calculations are supported with the C # programminglanguage application. The results of the job shop scheduling design with the threemethods, which method produces the best on minimization of mean tardiness.Thecompany job shop scheduling results on stamping parts products, there is a meantardiness of 61691 seconds. Based on calculations, active scheduling algorithmminimizes mean tardiness by 13.69%, % in the amount of 8443 seconds, non-delayalgorithm minimizes mean tardiness by 27.38% in the amount of 16894 seconds,and BPNN with greedy job alignment procedure minimizes mean tardiness by49.58% in the amount of 30587 seconds. The results obtained show that the bestmean tardiness minimization is found in the BPNN design with the greedy jobalignment procedure.