Enhancing Your Production Line's Effectiveness

Enhancing Your Production Line's Effectiveness

Rearrangement of existing resources (operators and buffers), without the need for any new investment in equipment or personnel is a cost effective way to compete in tough climates.

A primary concern for line managers has always been how to get the most out of their production system, given the limited amount of resources at their disposal. The present world-wide economic downturn has only emphasized the need to run the most efficient operations possible. Manufacturers everywhere are doing their best to cut overheads and enhance performance. This brings up the question of what managers could do to stay in business in this tough operating environment? One very interesting possibility to use current resources more effectively.

This article will discuss how managers can increase the efficiency of their plants by simply re-configuring the way they allocate personnel and material, rather than by making costly expenditure on new equipment.

Manual production lines are often made up of multiple stations serially connected. They play an important role in the world economy. Billions of dollars are spent yearly on their design, installation, operation, and repair. Even slight improvements in their performance can result in savings that, when calculated over the useful life of a line, are substantial.

Line Design Considerations

How you design your manual un-paced production line will substantially influence its efficiency. For instance, how to position workers who operate at different speeds, or vary their speeds in the course of the day, or where to keep work in process pieces along the line are just some of the issues.

Operator Average Processing Time

Since workers can work at their own pace and rhythm, mean service times (MT) are usually unequal, leading to an "unbalanced" line. Due to the nature of the product being manufactured, certain technological and precedence constraints might make it difficult to allocate work evenly among the stations, resulting in some stations having more work to do than others. Another source of imbalance is the operators themselves. As people have different levels of skill, competence, training and education, some are able to process items more quickly than others, further contributing to unequal average work times.

Operator Variability

Knowing mean service time values alone is not enough. You need to measure also the "scatter" or dispersion associated with these processing time averages. Research has shown that individual operators can vary in their work rate over the day by up to 66% from their average rate. A number of reasons account for this speed variation including fatigue, boredom and performing complex or changing tasks. In addition, people in general cannot perform a series of tasks differing in complexity and specificity again and again at exactly the same speed over a length of time. This is termed worker "variability". One measure of relative variability is the Coefficient of Variation (CV) -- the standard deviation of work times, divided by the average operation time.

Buffer Allocation

Because operators work at different speeds storage places, called "buffers", are usually placed in between stations in order to reduce the chances that a station is either being starved of work or is being blocked when it cannot eject a finished work piece to the succeeding station. One important decision is to determine the location and size of the buffers, since the unfinished pieces waiting around in the buffer stores represent tied-up capital that could be invested elsewhere. An important goal of "just-in-time" production is to maintain "lean buffering" i.e. keeping average buffer contents down to a minimum.

While unbalanced lines were previously viewed as being inefficient or undesirable, research has demonstrated that if the slower workers are placed at both ends of the line and the faster workers are positioned in the middle, output goes up. This is known as the "bowl phenomenon" because mean service times take on a bowl shape. Researchers have observed that the bowl phenomenon applies also to the variability of employees.

But what if unequal mean operation times, variability patterns, and buffer sizes simultaneously co-existed? Would it still be possible to achieve increased output and profitability from such unbalanced lines? Or would it be best to insist on trying to balance the line? We decided to put that to the test.

Our Computer Simulation Experiments

We simulated 5- and 8- work station lines, with different speeds, or mean times for the operators. The degree of imbalance (difference in the speeds between successive stations in the line) was set at three levels: a 2% difference (slight), a 5% difference (medium) and a 12% difference (high). Three CV values, representing steady, medium and variable operators, were used. The total buffer space allocated for each line was set at capacities of 8 and 24 units for the shorter 5-station line and at 14 and 42 units for the 8-station line (giving rise to average buffer sizes of 2 and 6 units in both cases).

Three performance measures were generated:

  • Line throughput rate
  • Overall average idle time percentage
  • Average buffer level of the whole line

Configurations We Examined
We studied several configurations of MT, CV and buffer capacity imbalance. With respect to their MT, workers were arranged in four different ways:

  • Key
    TR = throughput rate
    IT = idle time percentage
    ABL = average buffer level
    MT = mean service time
    CV = Coefficient of Variation
    The slowest worker first, followed by progressively faster workers -- pattern (\)
  • The fastest worker first, followed by progressively slower workers -- pattern (/)
  • The fastest workers placed at both ends of the line and the slowest worker(s) positioned in the middle -- pattern (^)
  • The slowest workers allocated to both ends of the line and the fastest worker(s) put in the middle -- pattern (V)

Four patterns of CV were looked at:

  • The most variable worker first, followed by steadier workers -- pattern (\)
  • The least variable worker first, followed by increasingly more variable operators -- pattern (/)
  • The most variable workers placed in the centre of the line -- pattern (^)
  • Locating the steadiest operators towards the centre of the line -- pattern (V)

Four configurations of buffer size allocations were considered:

  • Buffer capacity is concentrated at the front of the line -- pattern (\)
  • More buffer capacity is placed towards the end of the line -- pattern (/)
  • Assigning more buffer units nearer the middle of the line -- pattern (^)
  • Smaller buffer sizes are positioned towards the centre of the line -- pattern (V)

Best Pattern Found -- Output Rate and Idle Time

Our investigation has shown that the best configuration in terms of both increased output and lower idle time was the one where the fastest workers were placed at both ends of the line and the least variable workers were located in the middle (an inverted bowl for MT with a bowl shape for CV), combined with larger buffers assigned to the front of the line (/). Comparing our results to those of a balanced line counterpart, we found that for a line having 5 stations, an average buffer capacity of 2 units and a 5% degree of imbalance, an improvement in output of almost 3% was possible. In the case of a 5-station line with an average buffer size of 2 units and an imbalance degree of 2%, idle time was reduced by over 32%.

Figure 1 shows the best unbalanced pattern in respect of TR and IT for a line length of 5 stations.

Figure 1: The best TR and IT pattern for a five-station line.

Best Pattern Observed -- ABL

As far as ABL is concerned, we noticed it was best to have the slowest worker to man the first station, followed by progressively faster and faster workers, while at the same time the most variable (high CV) operators are allocated to both ends of the line and the steadier (low CV) workers are placed in the middle (a decreasing order for MT with a bowl configuration for CV). Combined with that, preference should be given to the end of the line, with the slowest and most variable workers being given larger buffer sizes. For an 8-station line with a mean buffer capacity of 6 units and a degree of speed difference of 12%, the above pattern resulted in a reduction in ABL of over 90% when compared to a balanced line arrangement -- an incredible degree of buffer content saving. Figure 2 shows the best ABL pattern found for a five-station unbalanced line.

Figure 2: The best ABL pattern for a line length of 5 stations

Generally speaking, our study has clearly shown that it is possible to get better performance with an unbalanced line, as opposed to a balanced line if the right joint imbalance pattern was employed.

Wise Decision Making

Unfortunately, we discerned no one pattern that simultaneously provided higher output, lower idle time and lower average buffer level. Therefore, managers need to make a difficult choice as to which effectiveness measures should be selected in their line design.

This decision will obviously rely on the type of operation and industry. In a sector with high demand and high labour utilization, such as on assembly lines in the household goods industries (fridges, washing machines, TVs, etc.), or where manpower is expensive, the relative cost of idle time will likely be greater than inventory keeping cost, and therefore you will make your decisions on the basis of minimising idle time and maximizing throughput rate.

On the other hand, inventory holding cost will likely be greater than idle time cost if it is more important to meet short lead times and adhere to lean manufacturing and just-in-time principles, as in the automotive industry, or in baked and other perishable goods production. In these cases your decision criterion will be based on average buffer level, so the patterns bringing it down would be the most suitable.

Benefits from Unbalancing Are Free

Studies have revealed that managers usually operate their lines by allocating workers and resources according to what has worked in the past and rarely resorting to expensive software or advanced analytical methods. We hope the guidelines listed here will be useful to you in the search for improved efficiency for your production line. Our results show that substantial improvements in performance are possible through unbalancing your line in an appropriate manner. However, our results also indicate that if one were to allocate resources in a haphazard manner, the consequence would be a dramatic decrease in line efficiency. The key point to remember is that the above gains can be thought of as "free,' since they only require a rearrangement of existing resources (operators and buffers), without the need for any new investment in equipment or personnel -- something that is critical in these trying economic times.

Dr. Sabry Shaaban is an Associate Professor, and Tom Mcnamara is a visiting lecturer at ESC Rennes School of Business in Rennes, France. http://www.esc-rennes.com/faculty/permanent-faculty/finance.php

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.