Skip to main content

Distribution Center Location

FACILITY LOCATION PROBLEM

Strategic distribution center placement reduces agricultural transport costs by 15–30% and post-harvest losses by 10–20%. In developing countries, 30–40% of fresh produce is lost between farm and market due to inadequate cold chain logistics (Parfitt et al., 2010). Every season, cooperatives must decide which potential hub locations to open—balancing facility costs against transport distances to dozens of farming regions. This is the Uncapacitated Facility Location Problem (UFLP)—one of the foundational NP-hard problems in combinatorial optimization.

Where This Decision Fits

Post-harvest operations chain — the highlighted step is what this page optimizes

Harvest Field collection
Silo Packing Storage allocation
Crop Transport Route trucks to fields
Distribution Center Hub placement & assignment
Market & Sales Pricing & contracts

The Problem

From distribution centers to optimization theory

An agricultural cooperative considers 5 potential distribution centers to serve 10 farming regions. Opening a center has a fixed cost; each region must be served by one center, incurring a transport cost. The goal is to minimize total fixed + transport cost.

This is the Uncapacitated Facility Location Problem (UFLP): select facilities to open and assign customers to minimize total cost. NP-hard.

Agriculture DomainUFLP Model
Distribution centerFacility (fixed cost fi)
Farming regionCustomer
Transport costAssignment cost cij
Minimize total costmin Σfiyi + Σcijxij
UFLP — NP-hard; greedy heuristics with local search

Try It Yourself

Design your own distribution network — edit centers and regions, then solve

Configuration

5 Centers · 10 Regions
Choose a Scenario
A regional cooperative with 5 candidate hub locations serving 10 farming regions. Moderate facility costs create a classic cost trade-off between opening more centers and longer transport distances.
Transport Cost Multiplier
2.0
Candidate Centers
NameXYFixed Cost ($K)
Farming Regions
NameXY
Facility Location Map

The Algorithm

Greedy Add for Facility Location

GREEDY ADD Start empty, open one at a time Step 1 Step 2 Final GREEDY DROP Start all open, close one at a time Step 1 Step 2 Final Opens best single facility first, then adds more until no improvement Starts with all open, drops least impactful until no improvement KEY DIFFERENCE They can find different solutions because the greedy path depends on the starting point. Open facility Closed facility Dropped this step Customer (farming region)
1

Start with No Facilities

All facilities are closed initially. Total cost is infinite (customers unserved).

2

Evaluate Each Candidate

For each closed facility, compute the total cost if it were opened (fixed cost + transport from all customers to nearest open facility).

3

Open Best Facility

Open the facility that gives the greatest cost reduction. Add its fixed cost to the total.

4

Repeat Until No Improvement

Continue opening facilities until no additional opening reduces total cost.

Real-World Complexity

Factors beyond the basic facility location model

Seasonal Roads

Some routes are impassable during spring thaw or winter, changing transport costs.

Capacity Limits

Real centers have throughput limits, making capacitated models more realistic.

Demand Growth

Regional production changes over years; locations must serve future demand.

Community Impact

Center locations affect local employment and community development.

Environmental Rules

Zoning and environmental regulations restrict where facilities can be built.

Multi-Modal Transport

Rail, road, and river transport options create different cost structures.

References

Key literature on facility location

Daskin, M.S. (2013).
"Network and Discrete Location: Models, Algorithms, and Applications."
2nd ed. Wiley.
Li, S. (2013).
"A 1.488-approximation algorithm for the uncapacitated facility location problem."
Information and Computation, 222, 45–58.

Need to optimize agricultural
distribution networks?

Get in Touch
Data shown is illustrative. This is a simplified model for educational purposes.
 Home
ESC