Supply Chain Simulation Optimization

A Mosaic Data Science Case Study

Download PDF



A leading electronic materials company that produces and distributes high-purity chemicals and gases required analytics consulting to optimize the way it supplies chemicals to one of its largest clients, a major international semiconductor manufacturer. In addition to supplying key chemicals and gases, the electronic materials company offers its clients supply chain and logistic services to improve manufacturing efficiencies, reduce costs, and increase revenues.

Due to the complexity of the semiconductor supply chain, the electronic materials company turned to Mosaic Data Science for expert supply chain consulting. Key objectives of the partnership were to improve customer service levels (i.e., delivering high-purity chemicals punctually) and to identify cost-efficient policies and procedures (i.e., optimizing inventory levels).

To achieve these objectives, Mosaic data science consultants developed a robust supply chain computer simulation for inventory policy experimentation and supply chain optimization.


Mosaic Data Science, a top-tier analytics company, developed a powerful supply chain simulation in Simio, complete with variable inventory policies for experimentation, a ‘pull’ system production process to accurately portray the in-place lean manufacturing strategy, and stochastic features to properly account for demand and delivery uncertainty.

Mosaic began the analysis with a supply chain architecture mapping effort. Here Mosaic came to understand the frequency and nature of product orders (i.e., the demand signal) at the various fabrication facilities around the world. This mapping effort included details of how the electronic materials company satisfied those orders through its worldwide network of warehouses, its international transportation logistics, and its product quality assurance and testing process.

Using an agile, customer-facing systems engineering process, Mosaic next began to iteratively design and develop a Simio computer simulation to implement the details of the supply chain system. After careful design elaboration and incremental software deliveries, the partnership established three Key Performance Indicators (KPIs):

  1. Stockouts (a count indicating how many times the customer ordered a product and it was not available; lower numbers are better),
  2. Delivery times (a duration capturing how long the customer waited for an ordered product; lower durations are better), and
  3. On-hand inventory (frequency histograms that gave the data analytics team keen insights into how well inventory levels in the warehouses vary in response to supply and demand fluctuations).

For added value, the Mosaic data science consultants designed the simulation keeping a keen eye on flexibility and scalability. In the supply chain business, inventory can make or break net operating margins, and in the high-stakes semiconductor industry, material shortages that shut down the fabrication line are not an option. Supply chain simulation, if done correctly, affords the opportunity to theorize not only new inventory policies but also new inventory warehouses that can increase supply chain reliability and robustness to unplanned disruptions while reducing overall system costs. And it does so without the high risk and extreme cost of an actual physical implementation. Taking advantage of Mosaic’s flexible and scalable design, the partnership hypothesized several new warehouse alternatives for evaluation with respect to the aforementioned KPIs.

Mosaic’s data science consultants also developed an automated analysis report that can produce comparison charts, summaries, etc. These go straight to the business customer to allow them to compare performance under different inventory policies.

Figure 1 depicts the simulation Mosaic delivered


Mosaic data scientists performed a thorough evaluation of the electronic material company’s supply chain through a careful experimentation regimen of various inventory policies, chemical quality levels, demand rates, and warehouse locations. For each scenario evaluation, Mosaic provided a suggested inventory policy and warehouse strategy, giving the electronics materials company more confidence in its supply chain performance and resilience. Mosaic’s advanced analyses and solutions created opportunities for increased customer satisfaction and loyalty. The customer was enthusiastic about the value of the analysis and simulation provided. They have since come back to request additional runs of the simulation to continue to receive value from the analysis and enhance their business relationship.

For More Information

Please contact Mosaic to learn more about this case study or our data-science services. Full case study available for download at the top of this page.