Hamilton Pozo
Any firm that is very interested in improvement faces a number of obstacles, the largest and most significant one being natural resistance to all and any type of change; this resistance is an inherent part of being human. In this context, one can see that:
- any improvement is a change; it is not possible to improve something without change; or as the saying goes - we cannot walk up the stairs without taking one’s foot off the previous step;
- any change whatsoever is seen by the majority as a threat to their security, given that it is unknown;
- any threat to safety causes emotional resistance;
- this emotional resistance delays the introduction of improvements.
These factors correspond to the climate perceived within the firm, which is the organizational culture. This provides solution to some of the problems of disintegration of organizations, by emphasizing the common ideas, ways of thinking, values, standards and ways of working (Freitas, 1991, Mu and Shen, 2007). However, organizational culture can also be a source of resistance to the process of change, of moving toward competitive advantage, making it stressful to carry out such change (Standard; Davis, 1999).
To soften this resistance, TOC diverges from the conventional method of learning, known as the Aristotelian approach. Conventional teaching produces faster results, but its effect does not last for long, which means that the results rapidly diminish. With this type of teaching, one tries to fight emotional resistance with logic. Whenever ones tries to overcome emotion with logic, emotion usually prevails. Successful learning is achieved when it takes place slowly, but continuously, as shown in figure 2, below (Goldratt; Cox, 2002). Therefore, one must arouse a strong commitment to change. One must ensure people enjoy the emotions of a creator, among other reasons because this will enhance their interest in ensuring the changes work out.
Figure 2. Socratic learning process.
Source. Goldratt and Cox, 2002.
As a result, the system TOC disseminates for making learning feasible is the same true and tried method of old, which has been proven to make people come up with ideas. Socrates employed this strategy when he wanted someone to do something, hence its name: the Socratic Approach.
Under this approach, answers are not taught. To the contrary, the person who wants to know something is asked questions and thus forced to discover or invent the answers. Asking Socratic questions is not as simple as it seems. It requires a special methodology and that the solutions are known in advance, broadly speaking. When the questions are poorly conducted, the outcome is merely the irritation of the interested parties. If everyone in the organization understands the firm’ s basic direction and its goals, then one has a common base and the people involved are more willing to ensure that transformations take place (Standard; Davis, 1999, Min-Shiang, Cheng-Chi Lee, Song-Kong and Jung-Wen, 2008).
In order to schedule activities with the aim of achieving the objectives, TOC assumes that it is first necessary to have a thorough understanding of the interrelationship between two types of resources usually found in every organization: bottleneck resources and non-bottleneck resources. Consider bottleneck resource X and assume that total market demand implies a monthly use of 200 monthly hours of this resource. In addition, as this is a bottleneck resource, it is assumed that demand is equal to the availability of this resource, in other words, 200 hours a month. By definition, the bottleneck resource is entirely utilized during the entire time of its availability. The demand for another resource Y, which is a non-bottleneck resource, comes to a total of 150 hours/month, although, like resource X, it has a production capacity of 200 hours (see figure 3).
Figure 3. Bottleneck and non-bottleneck resources.
Source: Authors’ adaptation.
With regard to the resources and the demands shown in figure 4, one may state that there are four types of relationships possible between these two resources.
The first of these is when all production flows from resource X to resource Y. In this situation, one can fully (100%) utilize resource X, but only 75% of resource Y is available time can be used, because resource X cannot produce enough to keep resource Y busy the whole time.
A second relationship, illustrated in figure 6, occurs when production flows from Y to X. Resource X will be used 100% of the time, and resource Y can be activated 100% of the time, as long as there is raw material available. However, bearing in mind that one of the TOC objectives is to increase flow and while also reducing stock and operating expenses, the conclusion is that Y should only be activated 75% of the time. Any activation, over and above this, implies in the build up of stocks in process, between resources Y and X, without any increase in flow.
The third relationship occurs when resources X and Y, instead of feeding each other, feed an assembly system that uses the parts through both. Resource X can be used 100% of the time. However, if resource Y is activated for more than 75% of the time, stock of this item will build up before it is assembled, since it is limited by the production capacity of resource X.
The last type of relationship that can involve the two resources is when they neither feed a common assembly system, nor feed each other. Instead, they feed independent market requirements. Once again, resource X can be used 100% of the time, but resource Y can only be used 75% of the time, otherwise it will lead to a build-up of stock of finished products. This is because demand remains limited and, to meet it, resource Y only needs 75% of its total processing capacity (150 hours a month).
Another assumption that TOC heavily emphasizes is the combination of the two phenomena previously mentioned: dependent events and statistical fluctuations. In other words, uncertain events will always occur in complex systems, such as production systems. The forecast level of demand is the basis for any firm’s strategic plan for production, sales, supplies and finance (Tubino, 2005).
Since it is difficult to anticipate where events will occur, all of the system is fragile or critical points must be protected. Moreover, the production of one item can entail a number of operations in terms of processing and transporting materials.
For most of these, the execution time varies according to a statistical distribution. In other words, the execution time of any given operation varies every time that the operation is performed. This implies that, in the production plan, when processing times (lead times) are used for a certain operation, in reality, it is the average lead times that are being used, which are subject to statistical fluctuations. (Slack, 2002) states that stock will occur whenever there is a difference between the pace or the rate of supply and of demand.
These fluctuations may arise from uncertainties in the operation, equipment capacity limitations, employee negligence, etc. No matter how much one establishes measures to control statistical fluctuations, it is impossible for production systems to eliminate the random component involved in the execution times of the several operations. Therefore, in all production processes, fluctuations exist to some extent, and they affect a substantial part of the operations of a process flowchart, if not all of them.
The fluctuations have roughly a normal distribution, given that it results from a series of random or uncontrollable events. If the operations involved in an item’s production process were not part of a sequence but were instead isolated, the sum of the fluctuations would tend to be zero. Delays in any particular activity would tend to offset other activities completed ahead of schedule, so that the deviation in the expected average time of execution would tend to zero (Goldratt, 2003).
However, manufacturing involves linking interdependent operations. Therefore, in this case, the chain’s statistical fluctuation does not average zero, as delays tend to spread throughout the chain. In other words, we do not have an average fluctuation, but rather an accrual of fluctuations. Moreover, in most cases, we have accrued delay, since the dependence limits the opportunity for greater fluctuations.
According to the logic of dependence between the linked events, TOC considers that the queue times depend upon how the scheduling is done. In fact, if a specific production order is given priority, for whatsoever reason, in a queue, waiting for a certain operation, this order will spend less time in the queue. As the queue time is one of the main components in items’ lead times, not surprisingly, the lead times will be different, in accordance with the scheduling of the orders. Consequently, if the lead times result from the scheduling, they should not be used as entry data for the scheduling process.
Thus, TOC approaches the problem in a different way, simultaneously taking into account the scheduling of activities and the capacity of bottleneck resources. Taking into account the capacity constraints of the bottleneck resources, the system decides to prioritize their occupation, and based on the defined sequence, calculates the lead times, which allows it to better schedule production.
For effective, optimum use of this theory, one should resort to its nine principles, which organize manufacturing management actions (Goldratt; Cox, 2002). These nine principles are set out below.
1. Balance the flow rather than the capacity.
The traditional approach is to balance capacity and then try to establish a gentle and if possible continuous flow of materials. TOC argues against balancing capacity and favors balancing the firm’s production flow. It is not capacity that should be balanced relative to demand. Instead, it is necessary to balance the flow of product through the factory with the market’s demand. The idea is to make the flow through the bottleneck equal to demand, since it is the first of these items that will limit the flow of the system as a whole.2. The use of a non-bottleneck resource is not determined by its availability, but by some other restriction of the system.
The use of the non-bottleneck resource should be determined by one of the system’s constraints, by the bottleneck resource or by market demand.
3. Use and activation of a resource are not synonymous.
There are crucial distinctions between using and activating a resource. Activating a non-bottleneck resource more than enough to feed a limiting bottleneck resource does not contribute to the defined objectives. To the contrary, the flow would remain constant though limited by the bottleneck resource. Meanwhile, the level of stock would rise, as would operate expenses, due to having to manage the ensuing stock. Since, in this case, the activation of the resource does not imply in helping the firm to achieve its targets, it cannot be called resource use, but only activation.
4. One hour gained regarding a bottleneck resource is a one-hour gain for the entire system.
The time available in a bottleneck resource is split between two components: processing time and preparation time. In the case of a bottleneck resource, if an hour of preparation time is saved, then an hour is gained in terms of processing time; in other words, the bottleneck resource becomes available for processing material. Moreover, one hour gained for processing in a bottleneck resource is not just a one-hour gain in the resource in question, but a one-hour flow gain for the entire production system, as it is a resource that limits the flow capacity of the system as a whole.
5. One hour gained in relation to a non-bottleneck resource is not a gain at all: it is just a mirage.
By definition, the time available of a non-bottleneck resource consists of three components: preparation time, processing time and idle time. Therefore, one hour of preparation time saved in relation to a non-bottleneck resource merely represents another hour of idle time for this resource, since the amount of processing time in the case of a non-bottleneck resource is determined, not by its availability, but by some other constraint on the system.
6. The transfer lot need not be and, frequently, should not be, equal to the processing lot.
In TOC, the transfer lot is always a fraction of the processing lot. This is the size of the lot that will be processed in a resource before it is prepared again for the processing of another item. The transfer lot meanwhile is the definition of the size of the lots that will be transferred to the subsequent operation. Since under TOC these lots are not required to be same size, amounts of processed material can be transferred to a subsequent operation, even before all the material in the processing lot is processed. This allows the lots to be split, enabling a reduction in the time that products spend to go through the factory.
7. The processing lot should be variable rather than fixed.
In TOC, contrary to what occurs in most traditional systems, the size of processing lots is a function of the factory’s situation and may vary from operation to operation. These lot sizes are established by the Theory’s calculation system, which takes into account the cost of carrying stocks, preparation costs, the flow requirements of certain items, and the types of resources, among others.
8. The bottlenecks not only determine the system’s flow, but also its stocks.
The bottlenecks define the production system’s flow, because they are the limiting factors for capacity. However, they are also the main factors that determine the level of stocks, because these have their volume determined and are located at points that can isolate the bottlenecks of statistical fluctuations spread by non-bottleneck resources that feed them. For instance, one builds up stock before the bottleneck machine, so that any delay does not lead to a stoppage of the bottleneck because of a shortage of material. This is achieved by creating of a time cushion before the bottleneck resource. In other words, the materials are scheduled to arrive at the bottleneck resource a specific amount of time before the instant at which the bottleneck is scheduled to go into operation.
9. The scheduling of activities and productive capacity should be considered simultaneously rather than sequentially. The lead times result from scheduling and cannot be assumed a priori.
As illustrated above, the queue times are a consequence of the scheduling and of the priorities scale. In this context, the lead times are consequences rather than assumptions.
[anterior] [inicio] [siguiente]
Vol. 32 (2) 2011
[Índice]