Bullwip effect

Let's imagine a workshop with 3 production stages, with following rules:

The initial inventory shows 3 parts remaining between cutting and welding, 2 parts remaining between welding and painting. No painted part is ready for next step, figuring the customer in this example.

A request is set for one blue part.

The requested blue part cannot be deliveried off the shelf, as none is available. It has to be manufactured.
Painting starts only to paint batches of 12. This quantity has been calculated as being the minimum batch size for painting operation, due to color preparation, cost of operation and tool cleaning. The paintshop having only two non painted parts at hand, welding has to provide at least the 10 other parts required to gather a painting batch.

Welding will process in batches of 7, that is the minimum quantity to start welding. Only 3 parts are available, so cutting must supply the four complementary parts.

Cutting always cuts 1O pieces, as it is its rule. Once cut, the pieces are pushed to welding. Welding has now 13 pieces ready. It start a batch of 7, in line with the welding rule and pushes it to painting once finished. Painting has now 9 parts available, but the batch size is 12.

Welding is required to weld another batch, but one piece is missing. The requirement is pushed upstreams to cutting. The latest will provide a new batch of ..10 pieces.

After welding finished the second batch of 7, painting has 16 parts ready and can process a standard batch of 12. From this batch, one part will be taken to the customer to answer the request.

Film of events

The film of events shows the wave triggered by the request going throught the process. The total of inventories varies from 3 units initialy up to 24 after the request has been fulfilled and the system stabilized again.

A small variation at the process end (delivery) triggers an amplifying variation all along the process upstreams called "bullwhip effect".

This phenomenon also demonstrates the Theory of Constraints motto, "Global optimum is not equal to the summ of local optima"

As a matter of facts, in our example an optimum batch size has been defined for each operation regardless to consequences on other links of this chain.

This example on an imaginary workshop with 3 operations happens also on large scale Supply Chains linking several entreprises.

This page was brought to you by ©hris HOHMANN -  http://chohmann.free.fr/