Skip to main content

SIM Model Building Process Overview

Understanding the overall process of building a SIM model, why a baseline is important and the outcome of building a model

R
Written by Renee Thiesing
Updated over 4 months ago

Building your first simulation model is a significant step toward validating your supply chain design and understanding how your network performs under realistic operating conditions. This guide walks you through the process of gathering data, populating the GAINS Simulation Excel template, and preparing your baseline model for execution.

Building a simulation model follows a logical sequence of steps. Understanding the full workflow first will help you stay organized as you move through each phase.

1. Plan Your Model Scope – Decide what portion of your supply chain to model: Which customers? Which sites? Which SKUs? What time period?

2. Gather Required Data – Collect the core data elements: customer master data, site information, SKUs, suppliers, and historical customer demand (orders).

3. Document Your Operating Parameters – Define how your network operates: inventory policies, transportation modes, sourcing rules, and operational constraints.

4. Populate the Excel Template – Fill in each table in the GAINS Simulation input template with your data.

5. Validate Your Data – Perform quality checks to ensure data is complete, consistent, and ready for simulation.

6. Upload to Architect – Import your Excel file into Architect to create a base scenario and prepare for simulation execution.

What is a Baseline Model?

A baseline model is a simulation of your supply chain as it currently operates. It uses historical data—customer orders, inventory positions, supplier lead times, and operational constraints—to recreate how your network actually performed during a specific time period. This is distinct from a scenario model, which tests hypothetical changes or strategies.

The primary value of a baseline model is validation. Once you build and run a baseline, you can compare the simulation's output to your actual historical performance. If they align, you've proven the model is accurate and can trust its projections for future scenarios. If they diverge, you've identified gaps in your understanding of your own operations—information that's invaluable before testing new network designs.

Did this answer your question?