USULAN METODE PENJADWALAN JOB SHOP DI PT X MENGGUNAKAN ALGORITMA SHIFTING BOTTLENECK YANG MENGAKOMODASI PENJADWALAN ULANG

• Main Author: Rafael Haryadi, Fransiskus
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/84157
• Institution: Institut Teknologi Bandung

PT X is a company that specializes in manufacturing custom machines for the automotive industry (job shop). Generally, PT X schedules its operations by prioritizing tasks with the nearest deadlines. However, this method is unsuitable due to the variation in the required number and duration of all the processes. This can risk lowering PT X's service quality in the eyes of its customers. Additionally, the overall scheduling is still largely manual, which takes a significant amount of time. This manual process is also applied to handling new jobs, where the schedule is adjusted one by one in a repetitive manner. Historical data supports the existence of this problem, showing that 37% of jobs are delayed overall. Thus, finding a scheduling model to help PT X manage new jobs is urgent. Given this need, the study began by reviewing potential alternatives for implementation. The alternatives included the shifting bottleneck algorithm, particle swarm optimization algorithm, genetic algorithm, and mixed-integer programming (MIP). Each alternative was evaluated to identify the most suitable algorithm for PT X. Based on the comparison results; the shifting bottleneck algorithm appears to be the most appropriate for addressing PT X's needs. Consequently, the shifting bottleneck algorithm was chosen for detailed testing with PT X's historical data. The results of testing the shifting bottleneck algorithm showed a reduction in makespan by 13%, from an initial 10446.2 hours to 9086.7 hours in a standard case. Additionally, the makespan for rescheduling, which initially took 12163.3 hours, was reduced to 8908.1 hours, a reduction of 27%. Therefore, scheduling with the proposed adjusted algorithm proved to have a positive impact compared to PT X's existing scheduling method.