PERANCANGAN LINTASAN PERAKITAN PRODUK EXCAVA 200 STANDAR DAN LONG ARM PADA DIVISI ALAT BERAT PT PINDAD (PERSERO) MENGGUNAKAN ALGORITMA SIMULATED ANNEALING

• Main Author: Irene Hutabarat, Theresia
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/77736
• Institution: Institut Teknologi Bandung

PT Pindad (Persero) is a state-owned defense equipment manufacturer. One of the main products of PT Pindad (Persero) is Excava 200 Standar and Long Arm. The company set production targets of 100 units for Excava 200 Standard and 50 units for Excava 200 Long Arm, but these targets have not been achieved due to poor operator assignment and task assignment methods. The desired cycle time is 600 minutes, while the current cycle time is 909,6 minutes. Therefore, assembly line balancing and operator assignment is required to minimize the cycle time in order to achieve production targets. The research was started by modifying the mathematical model to fit the actual assembly conditions. The results of the modification are the addition of 8 constraints to ensure one station can be assigned more than 1 operator, the same operation for both types of products is assigned to the same operator, and the upper assembly and lower assembly groups are assigned to different stations. Assembly line problems are an NP-hard resulting in very large computational time with analytical method. The cycle time obtained by the analytical method was 717.65 minutes for 168 hours of computing. A simulated annealing algorithm was developed for assembly line balancing with mixed products. The algorithm consists of two procedures: the outer loop and the inner loop. The outer loop executes the general simulated annealing procedure, while the inner loop generates initial solution and neighborhood solutions for the assembly line problem. The developed inner loop procedures include initial solution construction, task swapping, task inserting, and operator assignment. Simulated annealing algorithm produced viable solutions for the Excava 200 Standard and Long Arm assembly line problem with a cycle time of 128,14 minutes with 29 operators assigned to 12 workstations. PT Pindad will able to produce 500 units of Excava 200 Standar and 200 units of Excava Long Arm using this line and cycle time. Algorithm parameters were determined using factorial design method based on objective values and computation time. Experimental results indicated that the significant parameters for minimizing cycle time are the temperature reduction, iteration per temperature, and assignment repetition.