Calculating the Number of KANBANs

A supplier-customer relationship typically requires a cer­tain amount of stock to buffer each from the other's lack of regularity. In a totally regulated shop, no stock is needed, and no KANBANs are needed either. Demand is either perfectly level, or production flexibility is infinite: that is total Just-In-Time, and it does not exist anywhere. Like Total Quality, it is a journey. So while we do have stock, and we do want to manage it using KANBANs, this stock must be justified by KANBANs (Rules 3 and 4 order all inventory to be justified by a KANBAN). Each standard quantity of stock is assigned a KANBAN. How many KANBAN cards should be authorized? That is the ques­tion. How many KANBAN cards are needed to obtain an ideal in-process investment configuration which respects inter-operation imbalances and which protects an ad­equate level of operational efficiency? More cards in the customer-supplier feedback loop = more stock, more pro­tection, more costs, more lead time, etc.

Several formulas have been proposed to calculate the number of KANBANs required in a KANBAN loop. They all suffer from the same weakness: they consider each KANBAN loop individually, forgetting that efficiency and work in process (WIP) are rather a function of total mix and customer-supplier imbalances. This faulty reasoning is typical of most computerized approaches, even of MRP II systems, which ignore or simply report global impacts but rarely use them as a fine-tuning input.
For example, most of the formulas ignore setup times, which, any shop foreman will confirm, are, or should be, the most important factor determining lot sizing in a context of relatively fixed (short term) capacity. Resulting lot sizes force de-synchronized operations and flows that largely exceed unit demand by period, resulting in low stock turnover rates, inflexibility and long lead times.

If the KANBAN is to be a flow management technique, it must integrate these conflicts routinely. It must also integrate issues relating setup times to the volumes that must be produced and the current capacity constraints. We will borrow from the economic order quantity (EOQ) logic the notion of finding an optimum stocking configuration balancing setups with the cost of inventory. We will also use the planning tools available in MRP to establish the mix and forecast demand and capacity constraints.

The approach we propose to calculating KANBAN loops also borrows concepts from the LIMIT technique described by Haety, Plossl and Wight and documented by Fogarty & Hoffman (Ref. 3, p. 298). The purpose of this technique is to revise individual lot sizes when the lot sizing of all the parts needing to be produced exceeds available capacity. This is done using formulas which adjust lots sizes in
inverse proportion to the excess loading. The interesting aspect of this approach is really the idea of fine tuning individual lot sizes based on the global situation. The LIMIT technique has been used specifically to correct the cumulative effect of theoretical EOQs when the total capacity is insufficient to produce the lot sizes calculated.

We will use this approach, not to correct bad EOQs, but to directly determine lot sizes to adequately reflect the gen­eral shop constraints as a dynamic function of actual loading. These optimized lot sizes will then become impor­tant factors in establishing and managing the KANBAN. One major disadvantage of traditional EOQs is its well-known inability to adapt itself to the moving sets of constraints and opportunities being raised on the shop floor every day. The 7th rule eliminates this problem.

To validate that the KANBAN implementation proposed is acceptable from a managerial perspective, we will validate its results against stock turnover objectives as these are usually an important performance measure used to evalu­ate the Materials or Production functions.

A Few Basic Concepts

In the following sections, we will implement the KANBAN step by step. Keep the following concepts in mind as we implement an integrated solution using KANBANs.

• The shop is only as fast as its slowest mover, its bottleneck. It is wasteful to produce more on the other work centers. Spinning wheels uselessly only creates more in-process stocks and stretches manufacturing lead time. If plant throughput is insufficient, increase the capacity of the bottleneck first. Bottleneck condi­tions may move from one work center to another based on fluctuations in the mix.• Maximize the utilization of the bottleneck work center, of course.

• In the very short term horizon, the one relevant for the use of KANBANs, equipment and labor capacity are a fixed cost, and are therefore not considered a variable cost.

• Maximizing their utilization to create inventory (to appear efficient) only consumes raw materials too early, producing stock for which there is no immediate customer. See the second rule of KANBAN.

To implement the 7th rule, we must assess shop loading to identify work center constraints (bottlenecks) and oppor­tunities, and, as a result, to determine feasible turnover rates. Then, we will determine guidelines to set optimum lot sizes. We will determine the size of the KANBAN loops for each item produced. Finally, we will consider the KANBAN implementation and management process. The next 11 points and accompanying notes cover these issues.

To Be Continued

For balance of this article, click on the below link:

Lean Manufacturing Articles and go to Series 01