How to make production efficient

Efficient production (and indeed the enterprise in general) satisfies four simple requirements:

• Produces what the market needs
• Produces it with high quality
• Produces it quickly and on time.
• It does this at the lowest cost.

I will focus on two of four requirements here:

• how to produce quickly and on time
• how to produce at minimal cost

Goldratt's Theory of Constraints answers brilliantly to both of these questions. I have devoted the last few years to the study and implementation of algorithms . This is such an enterprise management system. First of all, a manufacturing enterprise. The effectiveness of which is noted even by such a "guru" of production management as Toyota.

It is impossible to put the whole such system in one post, but I will talk about the principles. And if someone considers it necessary to study it a little deeper, he will read the books of both Goldratt himself (“the very goal”) and Schragenheim (“production with incredible speed”). I am sure that reading these books will give you an answer to the seemingly rhetorical question “How to make production efficient.”

Why does it turn out that the “pure” time of production of a product and the real time of output vary by tens or even hundreds of times? Obviously, because there is a loss of time. They are of three types:

• resource waiting
• waiting for components (nodes)
• logistics losses

Imagine a production site. For example, a plot of painting. Any blanks that need to be painted constantly come to the site. The key question that needs to be answered is: “what to paint now, and what then” ?

If this is not understood, then the workpiece that needs to be painted right now will, as luck would have it, lie at the very bottom and will remember about it already when the customer fights in hysteria that his product has been overdue for a week.

And it will be the same in every section.

This is the expectation of a resource. The procurement lies in anticipation of the resource until it is free. But ... he is constantly busy with something else.

A system that clearly defines the sequence of tasks should be literally at every site of production. It should unambiguously determine which task should be performed in the first place, which in the second, etc.
The order of tasks is determined by the complexity of the task and the date the final product leaves production. As the release date approaches, the importance of the job grows, and it moves up the priority.

Here you have the product. It consists of three nodes (which also consist of something): U1, U2, U3. I will not draw, the artist of me is so-so. Moreover, in the video I did depict him on the board. The product must be manufactured 02/10/2013. Its final assembly needs an hour. This means that 09/08/2013. (the exact date is determined by a combination of labor input and production factors) in the final assembly area, all three nodes must be present.

The presence of the U3 unit on the final assembly site a month before is a completely undesirable phenomenon. This is worse than being late. Because, firstly, this will not affect the deadline for the finished product (there are no other two nodes), and secondly, this means that:

• used up components that could be spent on something really needed.
• place taken.
• you spent your money in vain (for the production of this unit). They had to be spent much later.
• resources were wasted in vain (for the production of this unit), and some NEEDED unit was waiting.

The ingenious Theory of Constraints not only talks about what you NEED to do at every moment, but also about what you DO NOT NEED.

The production of the U3 unit should be linked to the deadline for its delivery. That is, from 02/08/2013. Depending on the complexity of this node, a date is set before which this node cannot be produced. And starting from this date, the node changes color as the date X approaches. First it is green, then yellow, then red. If 02/08/2012 has already arrived, and it has not yet been completed, then it will be black.

At each site, all tasks must be performed in reverse order. First black, then red, then yellow, then green. White tasks should not be performed at all.

Waiting for components is an even more unpleasant situation that can completely disrupt all deadlines.
Take our notorious nodes. Morning 02/08/2013 The picker is in place. U1 and U2 too. U3 no. Where he was stuck - no one knows. The trial begins. It turns out that he is somewhere in the initial stage ... For the reason described above. So the production master will run, plugging such holes.

But waiting for the nodes is not so scary. Much worse is the expectation of purchased components. Because, as a rule, the purchase takes much longer than the production itself.

I won’t go into procurement management, because I wrote an article on this subject and even made a video . The main thing in procurement is to strictly observe the principles of priorities laid down in the Theory of Constraints. Always buy what you need and THEN when you need it. Saving, thereby precious working capital.

Pro loss in logistics mentioned in the video, and in LJ regularly at describing things. The ingenious Henry Ford, in order to avoid these losses, for example, came up with a conveyor. A hundred years ago. To avoid losses in logistics, it is necessary to minimize the loss of time during the transfer of products from site to site.

There is one more nuance. The production management system should be simple and sustainable. Resistant to external changes in the first place. I once wrote about such changes here . The simplicity (external, of course) of the Theory of Constraints lies in the fact that you should always perform the top task, that is, the highest priority. You can’t top, do the second from above. All.