SimWell Combines Arena Simulation with Linear Optimization to Improve a Supply Chain of Seasonal Products
This supply chain’s context
This company produces a family of finished products for which the demand is highly seasonal. Planning the production is a real challenge. It would be very costly to simply smoothen it throughout the entire year, since the volumes of products are huge and the storage and capital costs get very high.
Also, with the company having recently acquired one of its main competitors, they needed to study the combined supply chains as a whole, to see if any savings could be made by modifying it whatsoever. Thus, our goal was to model the North American East coast logistics in order to recommend the best supply chain structure, the best way to forecast the demand and the best general production strategy.
The tools used for modeling the demand and distribution
The first part of the project was a thorough statistical study of historical demand to understand the high variability of the seasonal demand for the products. The concept used to generate the demand in the Arena model is a Markov chain, which is often used in weather dependent situations. But when it comes to optimizing the production and balancing the distribution and storage throughout the company’s multiple sites, an optimization tool was needed. A software by Frontline Solvers ® was used for this purpose. A linear programming model was designed and used to determine most of the working parameters for the actual Simulation model.
Then, these inputs are imported into Arena where one year is simulated. At the end of the year, the final stocks are outputted in the linear-programming solver. These results are used to optimize a new yearly production plan, which is imported back in ARENA™, and so on. The project involved optimization, Markov Chain, Monte Carlo simulation and discrete event simulation skills.
Benefits for the customer
In the end, the project allowed to compare several multi-million dollar investments for improving the supply chain’s profitability and will help this company make decisions in the near future.