Mixed-Model Assembly Line Balancing in The Process of Assembling Trimming Area to Minimize Workstation Using RPW-MVM Method

. The development of an industry that continues to move forward coupled with global competition and openness demands that the company continues to evolve and always make improvements in improving the performance of its production process. XYZ Corp is an automotive company based in Germany that produces premium cars, the car assembly process groove at XYZ Corp is a trimming line, mechanical line and finishing line, XYZ Corp produces A-model, B-model and C-model cars. At this time XYZ Corp not achieving the production target due to the car assembly cycle time on the trimming Line 1 area exceeds the specified takt time. Assembly line balancing is required in the trimming area using Mixed-Model Assembly Line Balancing Problem (MALBP) approach to minimize the number of workstation, in the Trimming area assembly line balancing study using the Ranked Positional Weighted with Moving Target (RPW-MVM) method. Alocation constraint should be added due to machine restrictions that cannot be moved. After assembly line balancing, there was a decrease in the number of workstations to 14 workstations with a line efficiency of 86% and balancing efficiency of 97%.


Introduction
The development of the industrial that continues to move forward coupled with competition and global openness to meet consumer demand makes companies, especially automotive companies continue to grow and always make improvements. The development of the sales market is very competitive and oriented to customer satisfaction. For example, in the automotive industry, products are made based on several different models and features that are oriented to the customers wants. In an effort to fulfill the demand by building and maintaining assembly lines, the company made a single assembly line for several mixed models. In a situation like this, the problem of the balance of the assembly line will arise because of the demands for different models and features that vary with each product produced (Haq, Jayaprakash, & Rengarajan, 2006).
XYZ Corp is a German automotive company that produces premium cars with high performance. Cars manufactured by XYZ Corp has several types of car products such as model A, model B, and model C. The flow of the car assembly process at XYZ Corp is a trimming line and a mechanical line. While the engine, door, tire, and dashboard subassembly process is an assembly sub-process on the assembly line in the production department, according to Kumar and Mahto (2013) the assembly line model at XYZ Corp has the characteristics of a mixed-model assembly line because the company produces several types of products on 1 track and there is no setup time for changes in the types of products assembled.
The company can only meet the production target of 90%, the processing time for each work station in trimming is different, Figure 1 is a graphical representation of the processing time in the trimming area. Based on the graph data in Figure 1, it can be seen that at some work stations in the trimming area there is a processing time that exceeds the specified takt time, resulting in an unbalanced trimming assembly line which results in some work station bottlenecks and idle time so that the production target is not reached. Therefore it is necessary to balance the workload in the assembly area. This study aims to design a proposed assembly line with an optimal number of work stations with better line efficiency and balancing performance index.

Fig. 1. Time Distribution for Each Workstation
This study focuses on the imbalance of assembly lines. The problem of the imbalance of the assembly line is due to the uneven allocation of workload and the time at several stations that exceed takt time as found in Figure 1. In this study, the balancing of assembly lines is a mixed-model, where assembly lines can produce different models, with adaptability to demand with high levels of efficiency (Reginato et al., 2016). The main purpose of balancing assembly lines according to Eryuk et al (2008) is to minimize the number of work stations with predetermined cycle times or minimize cycle times with the number of work stations that have been determined.
Research "The Design of PC300 Type Hydraulic Excavator Assembly Line System with Genetic Algorithm Method" conducted by Andri Amir (2008) found an increase in productivity from 12 units/day to 13 units/day. Research conducted by Eryuk et al. (2008) regarding "Assembly Line Balancing in a Clothing Company", found that balancing assembly lines using the Ranked Positional Weight (RPW) method resulted in a more optimal number of work stations and line efficiency compared to line balancing using the method Probabilistic Balancing Line. Research conducted by Amen (2001) regarding the solution of assembly-line problems using the heuristic method, found that the heuristic ranked positional weight method has stable results when faced with problems in the real world. Balancing mixed-model assembly lines conducted by Reginato et al. (2016) examines and balances workloads with 7 product models using ranked positional weight with moving target (RPW-MVM) method, the results obtained after workload balancing is increased capacity, line efficiency and balancing efficiency.
Balancing mixed-model assembly lines conducted by Reginato et al. (2016) examines and balances workloads with 7 product models using ranked positional weight with moving target (RPW-MVM) method, the results obtained after workload balancing is incrThe development carried out by Reginato et al. (2016) developed the RPW method in solving problems in the assembly line development, solving assembly line problems using the RPW method which is purely regulated with predetermined cycle times, so that accumulated imbalances can produce assembly line performance. less well, in the RPW-MVM method, there is a limitation of moving target value (MVM) which is used as a reference in balancing and allocating work elements at work stations so that the assembly line is maintained more smoothly and provides a better performance value. eased capacity, line efficiency and balancing efficiency.
This study aims to determine the optimal number of work stations so that it can increase the value of line efficiency and balancing efficiency in the trimming area.

Assembly Line Balancing
Assembly line balancing Is a task allocation procedure, where tasks are distributed equally for each assembly work station so that the workload of operators at each work station is balanced (RAJA, 2015). Balance assembly lines are used generally to improve the assembly process for serial work stations. Assembly line balancing is a technique that is often used by industries that want to implement lean manufacturing systems (Kumar and Mahto, 2013). Measures of efficiency (operational performance) of assembly lines can be measured based on (Damayanti and Toha, 2012): 1. Total work stations, each work station will require operating costs, called operational costs. 2. Idle time is a time that is avoided in the context of the efficiency of the assembly line and will result in idle time costs. 3. Waiting time, waiting time in the process is calculated as waiting time cost.
The purpose of balancing assembly lines according to Scholl (1995) is, as follows: a. Reducing the number of work stations (m) with a predetermined cycle time. b. Reduce cycle time with the number of work stations that have been determined. c. Improve line efficiency. d. Look for a feasible solution with specified cycle time and number of work stations.

Ranked Positional Weight with Moving Target (RPW-MVM)
Solving line balancing problems on assembly lines using the pure RPW method is set with a predetermined cycle time, so that in allocating work elements to the workstation based on a fixed cycle time, this will result in allocating work elements at each workstation having accumulated imbalances which usually produces poor assembly line performance, to anticipate accumulated imbalances with fixed target limits / fixed cycle times the allocation of work elements must be allocated anywhere and anytime so that line balancing methods are developed on assembly lines based on weighting with moving targets or Moving-Target (MVM). Moving-Target (MVM) calculations on assembly lines are carried out at each work station and balancing assembly lines based on the number of work stations that are then Moving-Target (MVM) values are used as a reference for balancing the allocation of work elements at work stations. Moving-Target (MVM) calculations in the assembly line balancing process will make it easier to configure work stations by allocating work elements in a balanced manner to a predetermined work station (Reginato et al, 2016).

Performance Indicator
The balancing analysis of assembly lines generated from the RPW-MVM method used is as follows: 1. Line Efficiency (LE) Line Efficiency (LE) is an indicator to measure the use of assembly lines that will have an impact on aspects of economic evaluation.

Balancing Efficiency
Balancing efficiency is an indicator that measures the quality of the allocation of work elements to workstations which has an impact on increasing production levels.

Alocation Constraint
In the assembly process in the trimming area there are several allocation constraint, this limitation is due to machine restrictions that cannot be moved, some work processes or elements that have allocation constraint including: a. Work elements C22, C21, C20, C19, C18, C17, C14, C11, C12, C16, C15, C13, C23, C10, C2, C8, C7, C6, C1 must be done at station 3. b. K2 and K12 work elements must be worked on the same station. c. Work element A2 must be worked on station 1

Assembly Line Model
The following model assembly lines at PT. XYZ has the characteristics of mixed-models (Raja, 2015), namely assembly lines making products with several variants, where the operation of work for similar product variants for different variants. The measured performance indicators are: a. Number of work stations b. Line efficiency

Balance of the Existing Assembly Line
The following is the existing performance index information: 1. The number of existing assembly work stations is 15 work stations.
2. The assembly line working time is 25500 seconds/day, cycle time is 3662 seconds/unit, so the capacity is 7 units/day. 3. The assembly line efficiency is 81% and balancing efficiency is 90%.

Calculation procedure using RPW-MVM
The following are the calculation procedure steps using the RPW-MVM method to balance workload: 1. The first step is to make a combined precedence diagram for all work elements in each model that is worked on the existing work station, for details can be seen on Figure. 3. 2. Next, calculate the proportion of demand. In balancing assembly lines with many models, each model has a different processing time. The average processing time of each model needs to be defined by determining the proportion of demand.
pdm =dm/D (3) Where, dm is product demand in period p, where model m = 1, ... M; and D is the total demand for all models produced. The proportion of demand for each model, the proportion of demand for model A models is 43.5%, model B is 27.4% and model C is 29.1%. 3. Calculate cycle time (TC) / takt time based on total production demand. Takt time is 3187.5 seconds/unit, the allocation of work elements must be under takt time. (4) 4. Calculate the to allocate work elements to RPW-MVM because this method takes into account all models of a mix-model assembly line.

Conclusions
Based on the results of the calculation and reallocation of work elements for each work station in the trimming area, the optimal number of work stations is found in 14 work stations, with the proposed performance assembly line performance better than the performance index of the existing assembly line. A comparison of proposed and existing performance indexes can be seen in Table 1. After reallocation for each work element, the graph of time distribution for each work station can be seen in Figure 2.