TL;DR
- Problem: Given a set of customer orders and several possible warehouses, it is difficult to decide which facility should serve each customer or where new warehouses should be located without comparing the operational and cost consequences.
- My role: Product owner and designer currently defining the optimisation problem, decision variables, constraints, scenario workflow, and comparison experience.
- Decision: Treat warehouse configuration as a network-design problem with two connected decisions: allocate orders to facilities, then evaluate alternative facility locations and combinations.
- Outcome: An in-progress tool that lets users test warehouse allocation and location scenarios, compare results, and understand the trade-offs between distance, transport cost, and warehousing capacity.
- Technologies: Product design, optimisation, scenario analysis, operational software, logistics data modelling.
What it does
The Warehouse Facility Creator is a decision-support tool for warehouse network design. It is being developed around three practical functions.
1. Allocate orders to warehouses
Given a set of customer orders and a set of warehouses, the tool helps determine which warehouse should service each order.
The decision can take account of:
- distance between the customer and each warehouse;
- transport cost;
- the warehousing capacity;
- the operational consequences of assigning demand to one facility rather than another.
This turns an apparently simple allocation exercise into a constrained network decision. The nearest warehouse is not always the best warehouse if it cannot handle the order, has the wrong warehousing code, or produces a higher total transport cost.
2. Decide where warehouses should be located
The tool also supports the location decision itself. Given orders and candidate locations, a user can explore where one warehouse or several warehouses should be placed to serve the network.
This allows the analysis to consider questions such as:
- Which location gives the best service coverage?
- When does adding a second warehouse improve the network enough to justify its cost?
- Which customers become expensive or difficult to serve from a proposed network?
- How does the answer change when the available locations or warehouse capabilities change?
The aim is not to produce a single supposedly perfect answer. It is to make the trade-offs in a network decision visible and discussable.
3. Compare scenarios
The user can run different scenarios and compare their results. A scenario might change the available warehouses, candidate locations, order demand, transport assumptions, or warehousing codes.
Comparing scenarios helps users understand:
- the change in total transport cost;
- the distance travelled to serve customers;
- how orders are distributed between facilities;
- which customers or orders change facility;
- the effect of opening, closing, or repositioning a warehouse;
- the sensitivity of the result to different assumptions.
This is important because a network model is only useful when people can understand why its recommendation changes. Scenario comparison creates a way to challenge the assumptions rather than treating an optimisation result as an instruction.
The product problem
Warehouse network decisions combine customer demand, geography, facility capability, transport economics, and operational rules. These factors are often analysed separately, which makes it hard to see the system-wide effect of a change.
A warehouse that looks attractive by distance may be unsuitable for a particular order. A cheaper allocation may increase service distance or concentrate too much demand in one facility. A new warehouse may reduce transport cost while adding complexity or leaving other parts of the network underused.
The product needs to bring those considerations into one decision workflow without hiding the assumptions behind a single score.
Product decisions
Separate allocation from location design
Order allocation and facility location are related but different decisions. The tool treats them as connected stages so that users can first understand how a network serves demand and then test how changing the network changes those assignments.
Support judgement rather than replace it
The tool is intended to support a network-design conversation. Users should be able to inspect the assumptions, compare alternatives, and decide which trade-offs are acceptable in their operational context.
Optimise for explanation as well as calculation
A useful result needs to show more than a final allocation or location. It should help explain the distance, cost, eligibility, and network effects that led to that result.
Current status
This project is currently in development. The immediate focus is on refining the model for orders, customers, warehouses, candidate locations, transport costs, and warehousing codes, then building a scenario workflow that makes the comparisons easy to understand.
The next stage is to test the model against representative network-design problems and confirm that the outputs are useful to the people who would make or challenge these decisions.
What this project demonstrates
The Warehouse Facility Creator demonstrates a product approach to optimisation:
- define the decision before choosing the algorithm;
- keep operational eligibility and business rules inside the model;
- separate network design from the allocation decisions that follow;
- expose trade-offs instead of presenting false precision;
- give users a way to compare alternatives and apply their own judgement.
The value of the tool is not simply finding a mathematical allocation. It is helping people understand how warehouse structure, customer demand, transport cost, distance, and operational rules interact.