9 Planning scales in warehousing
“The best way to predict the future is to create it.” —Peter Drucker
Warehouse decisions occur at strategic, tactical, and operational levels. Each level applies to 2 connected scopes:
- The distribution network level, where warehouses are positioned within the broader supply chain and influence overall logistics efficiency.
- The warehouse level, where the focus shifts to internal design, layout, and operational strategies once a location has been chosen.
The chapter first examines warehouses in distribution networks. It then applies the 3 planning levels inside individual facilities.
Planning levels form a hierarchy. Decisions at one level constrain choices at other levels. Network changes affect warehouse design, and operating results may require tactical or strategic revisions. Planners therefore need to identify these dependencies and choose a sequential or iterative solution process.
The distribution-network discussion uses the multi-level framework from Hendriks (2009). The warehouse discussion uses the design dimensions from Gu et al. (2010) and Rouwenhorst et al. (2000).
Introduction
Warehouse and logistics plans operate under uncertainty. Longer horizons increase exposure to changes in markets, customer preferences, and technology. A strategic plan therefore needs a long-term direction and explicit conditions for revision.
Strategic planning uses the longest time horizon. Its purpose is to choose current actions under uncertainty, as illustrated in Figure 9.1. These actions prepare an organization for several plausible futures and can influence which future occurs.
Planning levels in logistics networks
At the strategic level, managers determine the number, location, and size of plants, cross-docks, warehouses, and other network nodes. These choices place capacity for several years and require high capital investment, so managers revise them infrequently.
At the tactical level, managers connect fixed nodes and organize flows between facilities. The planning horizon usually spans months. Decisions include transportation frequencies, inventory policies, production schedules, and the internal layout of each facility.
At the operational level, managers schedule vehicles, assign people, and route orders within the chosen network and facility design. Horizons range from days to seconds. These decisions control current work and respond to disruptions.
Planning levels inside the warehouse
After choosing a warehouse’s network role, planners decide its physical layout and operations. Gu et al. (2010) identifies 6 connected design dimensions: functional structure, departments, layout, sizing and dimensioning, operating strategy, and equipment selection. A change in one dimension can change the others, so planners must coordinate them.
Strategic decisions inside the facility
Strategic decisions inside a warehouse define the facility’s capabilities for many years and are seldom revisited. Examples include:
- Functional structure and footprint. The first step is to decide which functional areas the warehouse will contain (e.g., receiving, cross-docking, reserve storage, forward pick areas, value-added services, and shipping), and how much space each area will receive. These choices shape the fundamental flow of goods and constrain all subsequent layout and process decisions.
- Material handling system class. Next, planners choose the class of material handling system. Will the facility rely primarily on pallet racking and lift trucks, invest in automated storage and retrieval systems, or deploy conveyor belts, autonomous guided vehicles (AGVs), or other forms of mechanization? The level of mechanization sets the pace and cost of operations, and influences staffing and scheduling policies.
- Aisle orientation and block structure. Finally, one selects the global aisle plan and the positioning of docks. Choices such as a straight-through vs. a fishbone layout determine walking distances and congestion patterns. Once aisles and docks are set, later routing decisions must respect these constraints.
Tactical decisions inside the facility
Tactical decisions adapt the fixed shell to changing demand over a horizon of months. They do not modify the footprint but determine how to utilize it. Examples include:
- Detailed layout. Zoning the warehouse, assigning stock-keeping units (SKUs) to specific zones, and choosing a slotting policy family (e.g., dedicated, class-based, or random) are tactical tasks. The chosen slotting policy influences replenishment frequency and picker travel distances. Replenishment strategies for the forward pick area must also be tuned.
- Sizing and dimensioning. Within each zone, one decides the split between forward and reserve storage, the depth of lanes, and which SKU should occupy each storage mode (pallet rack, carton flow, shelving, etc.). Buffer capacities must be dimensioned to accommodate expected variability while avoiding excessive space.
- Equipment selection. Finally, planners decide whether order pickers move to the product (picker-to-part) or vice versa (part-to-picker), set batch sizes for order picking, choose between wave and waveless control systems, and define staffing plans.
Operational decisions inside the facility
Operational decisions enact the tactical plan on a daily or subdaily basis. Examples include:
- Scheduling. Warehouse managers schedule dock assignments, labor shifts, replenishment triggers, and cutoff times for order release. Scheduling must adapt to fluctuations in demand and supply, ensuring that capacity is available when needed.
- Routing and assignment. Once orders are released, pickers must be routed through the aisles, tasks must be batched, and replenishment assignments must be dispatched. Efficient routing reduces travel time and congestion, whereas poor routing wastes labor and delays orders.
- Real-time control. Warehouse management systems (WMS) handle exceptions and dynamically prioritize tasks. For example, a sudden stockout may require reprioritizing replenishment tasks, or a broken conveyor might divert orders to alternative routes.
Changing one design variable, such as the boundary between slotting classes, can alter replenishment volume, picker congestion, and dock schedules. Evaluate such changes with an operational model, such as discrete-event simulation or a validated queueing approximation, before adopting them.
Time scales
The planning horizons satisfy
\[ T_{\text{strategic}} \gg T_{\text{tactical}} \gg T_{\text{operational}}, \]
where \(T\) represents the time scale. Assigning precise values to these horizons, however, is not straightforward.
For instance, Rouwenhorst et al. (2000) suggests that for within-warehouse decisions, the strategic horizon \(T_{\text{strategic}}\) is about 5 years, the tactical horizon \(T_{\text{tactical}}\) is 2 years, and the operational horizon \(T_{\text{operational}}\) is 1 year. Other perspectives place tactical decisions on the scale of months or weeks, with operational planning unfolding over days, minutes, or even seconds.
Time scales depend on the decision context, requirements, and system complexity.
Coordinating levels through decomposition
A single model can combine strategic, tactical, and operational decisions. For large networks, this model often becomes computationally intractable, especially when operational decisions require real-time solutions.
Decomposition separates the planning problem into smaller, coordinated subproblems. Two strategies are common:
Sequential approach: The strategic problem is solved first, and its output becomes the input for the tactical problem. Tactical decisions are then fixed before addressing the operational problem. This method is fast and straightforward because each level is solved only once. However, it risks locking the system into a suboptimal configuration if the initial strategic or tactical choices are poor.
Alternating approach: Subproblems are solved iteratively. For example, tactical and operational models can exchange updated decisions. This process can improve coordination when a few interfaces control system performance. It uses more computation and may oscillate without a stopping rule.
In both cases, decomposition trades global optimality for tractability. Sequential planning suits stable environments with slowly changing data. Alternating planning suits high uncertainty or strong dependencies across planning levels.
- Use sequential planning when the environment is relatively stable and predictability is high.
- Use alternating planning when uncertainty is high or when a small number of critical interfaces dominate performance, so quick feedback across levels can significantly improve outcomes.
Conclusion
Warehouse planning connects long-term capacity choices, medium-term policies, and current operating decisions. Each level has a distinct horizon and changes the feasible choices at other levels. Effective plans use operating results to revise earlier decisions when their assumptions no longer hold.