From my doctoral course
This note comes from the doctoral course “Supply Chain Network Design”. I rewrote the explanations in my own words and recomputed the numerical example. The course’s main sources include Watson et al. (2013), Chopra and Meindl, and Dekker et al. (2004).
Supply chain and network
A supply chain is the set of suppliers, plants, warehouses, distribution centres and customers, and the flow of product (and money and information) from the source of production to the final customer. Network design (Supply Chain Network Design) is the answer to questions like:
- How many plants and warehouses do we want?
- Where should we build them and with what capacity?
- From which centre should each customer be supplied?
- Where should each product be produced?
The number and location of these facilities are among the most important success factors of a chain. Some experts say a large part of logistics cost can no longer be changed after the network is designed, since moving a plant takes years. (This figure is not exact; the qualitative point matters.)
When is design (or redesign) needed?
- Merger: after two companies merge, parallel networks become redundant.
- Expansion: entering a new market, growing demand.
- Disruption: a crisis, sanctions, a pandemic or a change of regulation has broken the network.
- Strategic change: for example from “low cost” to “fast response”.
Influencing factors
- Macroeconomic factors: taxes, customs tariffs, exchange rate, demand risk.
- Competitive factors: distance from competitors, competition for customers and labour (spatial competition games such as Hotelling’s model; see Static games: the prisoner’s dilemma and Nash equilibrium).
- Costs of disruption and the willingness to change.
- Logistical and operational factors: transport cost, inventory, response time.
The main relation: as the number of warehouses rises, transport cost falls (we get closer to the customer) but fixed cost and inventory rise (each warehouse needs its own safety stock). Total cost has an optimal point.
Types of facility
- Distribution centre (DC): sells to the customer or retailer from the warehouse.
- Warehouse at the plant and central warehouse (hub): the place of intermediate storage.
- Product mix: if a set of products must reach the customer, multi-product distribution centres are needed.
What models do we have?
Network design is solved with mathematical optimisation: basic location models, multi-product and multi-period models (multi-product, multi-period model), multi-objective (multi-objective network design), reverse logistics (reverse logistics and closed-loop chains), and models under uncertainty, risk and competition.
My note
I think network design is a good example of “a design whose mathematics is smaller than its decision”: the model assumes many things (constant demand, linear costs) and the decision is for ten years. So the model’s result must be tested with scenarios and sensitivity analysis.
Related: The multi-stage framework: from region to exact site.