[Total: 0    Average: 0/5]

The different Kanban have the same characteristics but their values or methods of calculation are quite different.


Whatever the type of Kanban, they all have the same characteristic: it is information given by the downstream workstation that gives the “right” to produce upstream.

However, depending on the Kanban, not all of them contain the same information and we detail below the characteristics of each.

The specific Kanban

The Specific Kanban is based on cards indicating the part reference parameters and associated quantities, which are constant. The whole issue is based on calculating the number of cards, determining the level of security stock between each workstation, necessary to ensure continuous production.

The size of a Kanban may be less than or equal to the production between 2 cycles of Mizusumashi ?  But this rule is not mandatory and does not fit all situations.  The number of Kanban must allow to cover the existing hazards in the system at the moment when it is put in place (adjustments, failures, non-quality…).

The number of Kanban is calculated via the following formula:

N = (1 + β) * (D * L)/C

  • D: represents the average consumption of products by customers per unit of time.
  • L: The time limit for the availability of products. This delay includes cycle time, settings and travel to bring the container.
  • C: The number of parts contained in a container.
  • 1 + β: It represents a safety factor that is determined according to the robustness of our processes: 1 We are sure of our processes, 2, we feel that our processes are not robust…

It will be held that generally, the Kanban must be dimensioned to us by respecting a minimum of safety stock which is calculated as follows:

Safety Stock = quantity of coins consumed on average during the supply time of the card + estimation of the supply variability of the card + variability in customer consumption

The generic Kanban

The Generic Kanban Adapts in real time to the production. The different parameters of this Kanban are constantly evolving. To calculate its characteristics, it is not possible to refer to a generic formula. The computer system that drives the production will generate the Kanban according to the need at the moment T.


The CONWIP Allows you to determine the optimum level of in-progress that can be used to produce a continuous flow based on demand. It is based on 3 parameters:

  • M: The number of cards and therefore the maximum stock of the line.
  • Q: The amount to be produced over the period.
  • CT: cycle time.

Statistical analysis of the process is necessary to optimize these parameters. It is carried out in several stages:


Put a number M of Cards.


Identify the desired productionQ over the period of time as well as the desired maxCT.


Make a statistical follow up on several equal time periods and calculate each time:

  • The average µ of the number of parts produced
  • The standard deviation σ of the number of parts produced
  • Average cycle timeCTavg



  • CTavg  > ct max : Review the value of Q and resume in step 2
  • Μ > Q + 3 σ: remove 1 card and go to step 3
  • Μ < Q-3 σ: add 1 card and go to step 3


With respect to 3 σ, it is considered that when a variation of the mean cycle time between the Position is greater than ± 3 σ, the process is said to be “Out of Control” (the same principle as the Control card).

Practice Advice

Attention, the formula for calculating the number of Kanban should be considered with great caution. Common sense is essential in setting up the Kanban.  S. Shingo indicates in his book “Control of production and Kanban method“:

« The way to determine the number of Kanban is not the most important. The important thing is to ask how should the production system be improved to fixer a minimum Kanban number? »

To answer that question, we’re going to have to:

  • Reduce the time needed to change tools
  • Reduce production times
  • Reduce machine failures.
  • Reduce the number of non-compliant parts.
  • Remove security stocks.


R. Guion, H. El Haouzi, A. Thomas (2011)-Study of the suitability of an adaptive Kanban with Multicriteria constraints: Case of a cutting cell

V. Gay, G. M. Bhargava (2007)-flow-through Production in a supply chain

N. Palekat (1999)-Job sequencing & WIP level determination in a cyclic CONWIP flowshop with blocking

W. J. Hopp, M. L. Roof (1998)-Setting WIP levels with statistical throughput control (STC) in CONWIP production line

Share This