
MTA vs MTO. Who will win? No one to anyone. We work
In my last article, I talked about how we are trying to put production flows in order by the theory of constraints. This time I will describe how we implement MTA (production to ensure availability).
Small input:
1. The project is closed. Designed for yourself.
2. The development environment does not matter. Algorithms of the theory of constraints are simple mathematics, which means that they are implemented by any available programming language.
3. All notes are an attempt to share our practice with the world. Look for the theory on the Internet.
The manager looked at how many products were in stock, and based on this, placed a warehouse order in the production plan. But the principle of doing something while there is free time has outlived itself. Let's say in stock 10 pcs of a certain product. The manager decides that this is not enough and tells the production: “We need to make 50 pcs.” While doing 50 pieces managed to ship half. Again the order for the warehouse. And so in a circle. The disadvantage of such a scheme is obvious. All data from the "Fonarev reference book". We decided that the workshop, although a living organism, should work according to easily controlled algorithms.
The manager selects the position that the system should monitor. Enters the source data. Only the size of the initial buffer is enough and this is the last thing a person does. Further autopilot. The initial buffer (see definition on the Internet) is what the system will push off to enter automatic mode. We will decide on the size of the initial buffer using the example shown in the screenshot below. The period is 12 months. Average stock 13 750 pcs. The turnover period is ~ 17 days. That is, in 17 days we will sell approximately 14,000 units. It is worth noting that these data are taken from the active database and at each point in time they are fresh. From practice, for comfortable work, we need to have a stock of about 1-1.5 months (depending on the product). It turns out that we need to keep approximately 20,000 units in stock. This is the size of the initial buffer. Precision is not important here. You can install 25,000 pcs. After a certain period of time, the system itself will adjust the buffer based on current shipments and arrivals to the warehouse.

All. Then the system starts tracking the positions we have chosen and notifies what needs to be placed in the plan when critical points are reached (red zone).

Blue is overproduction. Green is an excess. Yellow is comfort. Red is a danger. Black hole.

Next time I’ll talk about our dynamic buffer.
Small input:
1. The project is closed. Designed for yourself.
2. The development environment does not matter. Algorithms of the theory of constraints are simple mathematics, which means that they are implemented by any available programming language.
3. All notes are an attempt to share our practice with the world. Look for the theory on the Internet.
It was
The manager looked at how many products were in stock, and based on this, placed a warehouse order in the production plan. But the principle of doing something while there is free time has outlived itself. Let's say in stock 10 pcs of a certain product. The manager decides that this is not enough and tells the production: “We need to make 50 pcs.” While doing 50 pieces managed to ship half. Again the order for the warehouse. And so in a circle. The disadvantage of such a scheme is obvious. All data from the "Fonarev reference book". We decided that the workshop, although a living organism, should work according to easily controlled algorithms.
Has become
The manager selects the position that the system should monitor. Enters the source data. Only the size of the initial buffer is enough and this is the last thing a person does. Further autopilot. The initial buffer (see definition on the Internet) is what the system will push off to enter automatic mode. We will decide on the size of the initial buffer using the example shown in the screenshot below. The period is 12 months. Average stock 13 750 pcs. The turnover period is ~ 17 days. That is, in 17 days we will sell approximately 14,000 units. It is worth noting that these data are taken from the active database and at each point in time they are fresh. From practice, for comfortable work, we need to have a stock of about 1-1.5 months (depending on the product). It turns out that we need to keep approximately 20,000 units in stock. This is the size of the initial buffer. Precision is not important here. You can install 25,000 pcs. After a certain period of time, the system itself will adjust the buffer based on current shipments and arrivals to the warehouse.
All. Then the system starts tracking the positions we have chosen and notifies what needs to be placed in the plan when critical points are reached (red zone).
Blue is overproduction. Green is an excess. Yellow is comfort. Red is a danger. Black hole.

Next time I’ll talk about our dynamic buffer.