The 2 applications presented here can be accessed through Taipy’s website at:
- 2-echelon supply chain
- 3-echelon supply chain

Before we delve into practical examples, let's introduce some key concepts behind supply chain network optimization.

The concept of Echelon

Supply Chain models differ widely in their complexity. One such level is the concept of Echelon or the number of layers required to model the actual supply chain.

If we take the example of a country-wide pharmaceutical wholesaler, it may only be interested in optimizing the location of its own distribution centers. In such cases, only two levels or echelons need to be represented:

- The Distribution Centers
- The Customer locations

In the case of a manufacturer, an additional echelon might be needed: Plants.
Thus, a 3-echelon supply chain would include:
- Plants
- Distribution Centers
- Customer Locations

Similarly, a nationwide wholesaler might model a 3-echelon supply chain if they have multiple distribution levels:
- Regional Distribution Centers
- Local Distribution Centers
- Customer Locations
While it's possible to have even more echelons, in most real-life situations, 2-echelon or 3-echelon supply chains cover the majority of use cases.

Modeling various objectives

When optimizing supply chains, it is crucial to model one or several objectives.
Transportation costs are a primary objective. In a 2-echelon supply chain, the transportation costs relate to:
- The distance traveled between warehouses & customer locations,
- The mode of transport contract: FTL, LTL, etc.
- Specifics of transportation contracts, etc.

Advanced models also account for warehouse fixed and variable costs. Additionally, costs often include the carbon footprint generated.

Supply Chain Network Constraints

For certain use cases, it is paramount to model important constraints such as:
In certain scenarios, it is essential to model key constraints, such as:
- Maximum number of warehouses
- Capacity at different echelons like warehouse or plant capacity
- Plant capabilities: some plants can only produce certain products while others can produce a full range
- Pre-defined locations
- Maximum distance/time to delivery: ensuring a maximum distance/time between the serving warehouse and each customer
- Etc.

What is Supply Chain Optimization?

With core concepts defined, supply chain software typically employs an optimization model to generate “optimal” solutions for various scenarios.

Applications can be categorized into three different maturity stages:

Stage 1:
- Provides visuals to display input data
- Allows end-users to make manual decisions, such as selecting warehouse locations

Stage 2:
- Similar to Stage 1 but adds the capability for users to create and compare different scenarios/solutions

Stage 3:
- Fully functional Decision Support System
- Includes a smart optimization engine that generates optimal solutions
- Allows end-users to modify input parameters, automatically generate optimal solutions, and compare and select the best option
This later stage incorporates all the functions of a Decision Support System.

Examples of level-3 Supply Chain Solutions

Optimal Solution to a 2-echelon supply chain

Optimal Solution to a 3-echelon supply chain

Comparing Scenarios

Being able to easily compare scenarios is an integral part of Taipy.

Here’s an example of a 2-scenario comparison:

When more than two scenarios need to be evaluated, here’s an example of a multi-scenario comparison:

Conclusion

This article demonstrates how to build powerful Decision-Support Systems using Taipy, all in Python. The combination of powerful graphics and scenario management makes Taipy a unique tool in its category.