Rethinking Warehouse Design
The question we asked at Hanes Erie, Inc. was, “What must we do in warehouse design to make a significant impact on practice?” To answer this question, we set out to establish a Warehousing Roundtable. The Warehouse Roundtable would meet on a fairly regular basis—approximately quarterly—to discuss problems, practices, and results.
Conceptual Framework
In this framework, a function has a corresponding area and/or equipment in the warehouse where the function is performed. Thus, specifying the function involves specifying the technology to be used, and eventually, the details of the implementation, i.e., the configuration of the technology and perhaps associated control functions.
Flows between warehouse functions correspond to warehouse operations. For example, put-away is an operation that moves material from the receiving function to the storage function. Order picking is an operation that moves material from a storage function or pick-line function to a sortation or accumulation function. An operation may require equipment; for example, lift trucks, order picking trucks, conveyors, or other equipment. It also may require other resources, such as labor or computer systems. In addition, operations will require a method, i.e., a system of planning, managing, and controlling the operation. For example, an order picking operation may be specified to use single order picking or batch picking with picking zones.
At the conclusion of this project, we will have two important results.
Profiling
The term “profiling” is used to represent the process of developing a description of the warehouse requirements which incorporates the relevant characteristics of both the items to be handled, and the orders to be filled. In the ideal situation, warehouse design begins with a SKU master, a database describing each item in detail, and an order master, a database with a representative sample of customer orders for, say, one year. These two databases could be sorted, queried, partitioned, and statistically analyzed to answer such questions as:
- Can the items be grouped according to size? Weight? Type (i.e., flammable, etc)? Frequency of orders? Size of orders?
- Can the orders be categorized according to number of lines? Size? Weight? Value?
- Can the customers be grouped according to frequency of orders? Value of orders?
System Design Decision Optimization
In our warehouse design, there are many trade-offs that should be considered. More automation may reduce the labor cost, but increase the investment cost. More storage zones may improve the space utilization, but require more control. Zone picking may improve the productivity of the stock selectors, but requires more planning and control of picking. Trade-offs like these abound.
Ideally, it would be possible to formulate an optimization model that would consider the trade-offs between all the available options and select the decisions leading to the “best” solution, according to the specified criterion. This optimization model would enable us to deal explicitly with customer service levels, either as a constraint or as part of the criterion, which means it would require a formal definition and computable metrics customer service. Such an ideal formulation may be possible some day, but for the time being, we need more pragmatic approaches.
Consider, for example, the design of three key functions in the warehouse—the reserve storage, pick-line, and sortation/accumulation functions. Under what assumptions can we formulate each functional design problem using consistent definitions? If we can create “uniform” formulations, can we then identify a relatively small set of parameters, decisions, or performance values that can be used to coordinate the solutions of the three functional design problems?
These are the questions that motivate our current efforts on warehouse design decision optimization. We are seeking to develop a heuristic procedure that will allow us to coordinate the (sub) optimization of a set of functions in the warehouse.
For further information contact:
Peter C. Berry, PH.D.
Director of Quality and Sales
Hanes Erie, Inc.
Peter Berry
(814) 474-1999