Intelligent Manufacturing € February € 1997 € Vol. 3 € No. 2


Whirlpool Shifts from Push to Pull


Until recently, Whirlpool Corp. (Benton Harbor, Mich.; http://www.whirlpool.com), a manufacturer of major home appliances, relied on "push" manufacturing, a system of producing as much inventory as possible and pushing it out to the distribution channels. While this system allows plants to operate at capacity, it has the disadvantage of creating periodic oversupply and consequent discounting pressure in the distribution channels. Further, since production schedules had a weekly-based time step, a minimum four weeks' delay was required to fill any nonstandard orders.

In seeking to improve production planning, Whirlpool's goal was to reduce inventory, work-in-progress and product obsolescence - in short, to fundamentally alter Whirlpool's manufacturing model from push to pull. "Pull" manufacturing predicates factory production on actual customer demand, and therefore requires both very timely relay of orders to the plant, and a highly optimized production scheduling process to fulfill those orders. To achieve all this, production schedules had to be slashed from a weekly to a daily bases to reduce response time to customer order from four weeks to five days.

Whirlpool production planning is based on a statistical application called DRP (Distributed Resource Planning), which generates 52 weeks of unconstrained production requirements. To adapt DRP to the demands of pull manufacturing, Whirlpool developed an application called RAD (Reality Applied to Distributed Resource Planning).

The RAD project is based on ILOG Solver from ILOG (Mountain View, Calif.; http://www.ilog.com), which optimizes production planning based on pull manufacturing constraints. ILOG Solver is a C++ library for enterprise resource planning. DRP requirements are loaded into the system via graphical user interface by the master scheduler at each plant on a daily basis. The output from RAD is then passed along to the MRPII system which handles bill-of-materials, vendor parts releasing, shop floor scheduling and so on.

The RAD application has enabled pull manufacturing at Whirlpool. Appliances are built according to actual customer demand, with response time reduced from four weeks to five days. Inventory on display in dealer showrooms is automatically replenished when depleted by customer demand. Production costs, work-in-progress and finished-goods inventories have been greatly reduced, while responsiveness to customer demand has been much improved.

The RAD system is currently supporting eight Whirlpool manufacturing plants. It helps Whirlpool more accurately prioritize production, create achievable production schedules which leverage available plant capabilities, and monitor unsatisfied requirements resulting from production constraints.


Click here to return to Table of Contents for the current issue of Intelligent Manufacturing.

Intelligent Manufacturing Copyright © 2020 - Lionheart Publishing Inc. All rights reserved.