You are on page 1of 9
5a ater en Pio} 2. Assign a worker to the succeeding stations in the production ‘Sequence if the input to these stations is sulficient to meet the target Production for the period. We leave it as an exercise to the reader to develop the daily labor schedule for the example in the preceding section. Assembly Line Balancing ‘An assembly line is a method of production in which the parts are assembled and made into the final product as the unit progresses from station to station. To develop a balanced assembly line requires logical planning and involves the distribution of the total job among the work stations so that all stations can complete their designated tasks at approx- imately the same time, Ifthe line were perfectly balanced, the time required in each station would be identical. However, such balance is rarely possible, and the longest station time dictates the actual cycle time for the entire assembly, Various techniques are available for developing an assembly line, but the basic requirements for all are the same: break the total job content into small tasks or work elements and determine the precedence relationships, that is, decide which tasks must be performed before the next one begins. These relationships also indicate the tasks that can be performed simultaneously. For example, in paintbrush manufacturing, glue must be poured into a clip before the bristles can be inserted; however, at the same time, someone (or a machine) could polish and stamp the handle. Next, determine the time that will be required for each work element. Any other restrictions such as the requirement that none of a group of work elements may be placed together in the same work station, called negative zoning, or the requirement that some tasks must be placed in the same work station, called positive zoning, should also be considered and clearly stated. Mathematical methods are available to balance assembly lines in terms ‘that are mainly the application of optimization techniques such as integer programming and branch-and-bound methods. However, these exact ‘methods are very tedious and cumbersome to apply to a large assembly line ‘containing a multitude of tasks. We will show two heuristic methods that ‘are popular because of their simplicity. Balancing Procedures The methods are illustrated by applying them to the following sample problem. The elements are represented in Figure 5.13 bby nodes, and the precedence of each is indicated by the arrow connecting the nodes. In the diagram the clement (task) times are listed above their cor- responding nodes. The same data are also presented in Table 5.12. ‘The total time needed to complete all tasks is 5.31 minutes. bas eet 4, List the tasks in decreasing order of magnitude of task times, the task requiring the largest time being first. Also list the corresponding immediate predecessor task(s) for each. 2 Designate the first station in step 1 as station 1 and number the remaining stations consecutively. 3. Beginning at the top of the task list, assign a feasible task to the station under consideration. Once the task is assigned, all reference to it is femoved from the predecessor task list. A task is feasible only if it does not have any predecessors or if all predecessors have been deleted. It may be assigned only if it does not exceed the cycle time for the station, ‘and this condition can be checked by comparing. the cumulative time of all the jobs so far assigned to that station, including the task under consideration, with the cycle time, Ifthe cumulative time ee Cn oe eee eas 2 022 a : 042 ‘ 048 ; O12 ; i : ; be 063 sz 034 80 070 Wv Toble 5.13, ‘Task time eee o ater than the cyele time, the task under sais cannot be igned to the station. If no task is feasible, proceed to step 5. 4. Delete the task that is assigned from the first column of the task list I the list is now empty, go on to step 6; otherwise, return to step 3. 5. Create a now station by increasing the station count by one. Return to step 3. 6. Al jobs are assigned, and the present station number reflects the number of stations required. The procedure also shows the job ‘assginments for each station. The largest of the cumulative times for the individual stations is the true eycle time. We will demonstrate application of the above procedure to the sample problem presented in Figure 5.13 and Table 5.12 with a planned cycle time ‘of one minute or with a production rate of 400/day, assuming that there are 400 working minutes in a day shift. In the following the first step we list the tasks (elements) in the order of decreasing task time (Table 5.13). Also listed are the immediate predevessor(s) of each task. This information is needed in step 2 and can be obtained cither from the data table or from the precedence diagram. ‘The procedure begins with the element listed first in Table 5.13, task 10. ‘The accumulated task time would be equal to 0.80, which is less than the cycle time; but since neither task 7 nor 8 is completed, task 10 cannot yet be considered. The same procedure is repeated for the next element in the list, task 11, with the same result. As we go down the list, task 1 is the first feasible task to be encountered, and it is assigned to the station list. We delete all reference to task 1 in the predecessor list. In reality these operations are performed on the same table; but for clarity, we show the results in Table 5.14, ‘Next we return to the top of the list; that is, 10, 11, .... Task 2 is the next feasible task found with a cumulative time of 1.0 or less, namely, Task Task Time Immediate Predecessor(s) 10 080) " 070 2 070 7 063 8 ase 5 oss. 3 042 4 9 1 6 ~ 038 034 022 an Table 5.15 Final ask assignments 063 35 otal worker time required/unit) — job time/unit Total worker time required/unit ‘and is known as balanced delay. 54 Labor Assignments mm 4 Table 5.16 Ranked posit 2-4 Task RPW Task Time _Predecessor(s) 1 531 a — 3439 a2 1 2 37 ~~ 070 1 S307 as 23 i 7 247063 35 4 240038 13 8 2022082 4 wo 15080 ne 6 16 a2 5 9 10s os 67 070070 9,10 ay, In this case, seven stations are utilized, each worked by one person, station 1. with a cycle time of 0.92 minute, Therefore, efficiency is Pccontins soe eecennast pruk 3 bas 1 = [SOS 831) ,. 100 forte tine: c=!" 7x092 |” = = 82.5% os Procedure 11 — Ranked positional weighted method 1n the previous pn method the tasks were listed in decreasing magnitude of time requirements. + Jn this method they are ranked according to their importance to the ' ‘completion of all the tasks that depend on them. The importance i ‘ 1 measured by the ranked positional weight (RPW) of each element, which is the sum of the times for all elements that directly follow it in the precedence i diagram plus the time for the particular task itself. For example, the RPW for clement 5 is the sum of the times for elements 6,7, 9, 10, and 1 plus the time for element 5— that is, 0.12 + 0.63 + 0.34 + 0.80 + 0.70 + 048 = 307 } Once the elements have been listed in descending order of RPW along t with their predecessor elements (as shown in Table 5.16), from there on the ) procedure for developing an assembly line is exactly the same as before (steps 2-6). i 4 ‘Carrying out steps 2-6 of Procedure I (largest candidate rule) results in } the final assignments listed in Table 5.17. | Seven stations are still required, but the cycle time is increased to 098 j i ‘minute, the cumulative time associated with station 4. The efficiency of this ‘arrangement is , en [OPES 109 7x 098 =T14% q ‘The present production rate is 400/0.98 ~ 408 units/day. CHAPTER 5 Production Charts ond Systems Rev Final tx (gh stations Pourtesy of sa Table 5.17 RPW assignments Table 5.18 Final task assignments Leber Assignments 173 Cumulative Station Element Task Time Task Time 1 1 022 022 3 a2 064 2 2 070 070 3 3 0.48 048 4 038 0.86 6 042 98 4 7 0463 063 9 034 087 5 8 032 032 6 10 020 0.80 1 uw 070 070 It should be noticed that these heuristic procedures develop task assignments that might not be very satisfactory, as is the case here with an efficiency of 77.4%. One way to increase the efficiency is to try the procedure again with a different cycle time. For example, the cycle time of 0.9 minute results in the task distributions shown in Table 5.18, With an efficiency of 88%, the production rate is now increased to 400/0.86 ~ 465 units/day. It is also possible to increase the cycle time, thus reducing the production rate but increasing the efficiency by decreasing the number of required stations. Parallel Grouping of Stations In some instances the required production rate can necessitate having a cycle time that is even less than the time required for the completion of one of the tasks. For instance, in the previous case, if the output needed were 800 units per day, the cycle time would have to be 400/800 = 0.50 minute. Tasks 10, 11, 2, 7, and 8 require more time than that for their completion. Elements Assigned — Cumulative Station 10 the Station __Station Time 1 064 074 0.86 oss 0.86 0380 070 7. 8 (CHAPTER S Figure 5.14 Parallel stations Production Charts and Systems tance the previous methods, in which the stations were oper- Ated in series only, would have to be modified to allow for a parallel setup. When two or more stations are in parallel and all are assigned to perform the same tasks, the permitted time for completion of the tasks in a station is longer than the cycle time. As an example, suppose there are two stations in parallel, as shown in Figure 5.14(a), each requiring time ¢ to perform the tasks assigned to it; then two units are produced per ¢ time interval. Other stations in series may therefore operate with 1/2 as their cycle time. Similarly, if there are n stations parallel to each other, the eycle time for the remaining stations could be t/n, where the cycle time for the stations in parallel is t ‘A question might be raised as to why we do not have six parallel stations in our example each performing all operations. This arrangement should obtain production rate that is better than the one per minute needed and at the same time would reduce the labor requirement from seven to six. There are several reasons why such a grouping may not be possible. An additional task assigned to a work station might mean ‘additional investment at that station for the machines and tools that are necessary to perform the task. In multiple stations the same group of equipment would have to be available at each station, an investment that could be prohibitive. The task time could inerease because as the work area is enlarged, the worker must move farther to perform more tasks. Workers with more experience and training would be needed to perform all tasks. A larger floor space would generally be required if the same tasks were to be ‘epeated in each station, g 55 compute Progam Dserpten o It should be recognized that in a practical layout of an assembly line, two parallel stations need not be physically parallel. They can, for example, be two successive stations fed by a common conveyor. It is possible to dedicate alternate units on the same conveyor to cach of the stations in various ways, from using a sophisticated mechanical, pushj/pull arrangement to simply marking alternate positions on the conveyor by differently colored carriages such as trays. Computerized Assembly Line Programs Numerous computer programs have been designed for assembly line bal- ancing. In most cases, each was developed for a particular situation, but all are based on one of the following techniques. + Enumeration of all possible sequences + Selecting the best from randomly chosen sequences + Selecting a sequence that is similar to one already used in the company - Applying a mathematical programming technique such as linear programming, integer programming, or the branch-and-bound method + Using heuristic rules, such as the two we have just seen + Applying rules that are specifically developed for the problem, pos- sibly including among other items, restrictions on tasks and assign- ments among the tasks Applicable Programs by Others It is beyond the scope of this chapter to go into the details of all known computer assembly line balancing programs. The following are some of the ‘more popular. + NULISP: Nottingham University Line Sequencing Program, devel- ‘oped by Nottingham University in 1978 + ALPACA: Assembly Line Planning And Control Activity, devel- oped by General Motors in 1967 + GALS: Generalized Assembly Line Simulator, developed by ITT Research Institute in 1968 + COMSOAL: Computer Method of Sequencing Operations for Assembly Line, developed by Chrysler Corporation in 1966 Computer Program Description Four programs for this chapter are included on the program disk. For assembly line balancing the programs include one on the largest candidate rule and one on the ranked positional weighted method. The third and

You might also like