Summary
This multi-billion dollar, globally integrated mining and chemicals company uses River Logic's Enterprise Optimizer® (EO) Decision Engine for its mid-term and long-term planning.
The monthly planning solution looks across a 24-month horizon and shows which products are most profitable on a per-product and/or per-customer basis. The long-term planning solution supports yearly strategic planning across a 30+ year horizon. It helps the company understand, year by year, the best way to get its starting products out of the ground, given the existing mine plans its geologists have compiled and future market predictions. The initial implementation is valued at a $45-$80M Net Present Value (NPV) improvement, realized within the first 6 months of deployment.
Company Overview
The company operates globally, with almost two dozen mining sites and processing facilities and thousands of employees. It mines three naturally occurring minerals that are processed into its primary product. The primary product has around 20 code variations, determined by purity, size, weight, and other chemical characteristics. The company also sells secondary products.
Trade-offs Between Upstream and Downstream Processes
Mined ore varies on multiple attributes: concentration, chemical form, impurities present, and more. How the ore is processed depends on those attributes. There can be variations in blending, additional steps, recycling, removal of precious minerals, and so on. The minerals are used as feedstock either in their natural form or in an upgraded form produced through secondary processing. They are then loaded onto large barges and shipped to chemical processing plants to make end products. The company also has to decide whether to further purify and sell by-products.
What comes out of the mines, and the processing options at each mine, affect the cost and efficiency of final production and the potential end products. The company needed to account for upstream decisions (processing at the mine sites), downstream decisions (which end products to produce), and shipping variables together in order to optimize its entire value chain. It needed to see when it was better to de-optimize a process upstream in order to realize net improvements across the whole value chain.
The Challenge
The company's decision-making is multifaceted. Because it mines its own feedstock, it can either sell the feedstock or move it further along the manufacturing process.
The company must handle variations in the purity of starting products (feedstock) while weighing the logistical and financial impact of selling and further purifying by-products, and of whether to buy, sell, or continue certain quantities of feedstock in the manufacturing cycle. Some of the mined product is processed near the mine to offset transportation costs and increase the purity of the shipped product. So there are trade-offs in deciding which products and quantities get processed, where they get processed, and how products are transported to the next plant or end customer.
All of these factors made for an extremely complex planning process, full of significant trade-offs, variables, and constraints that need to be evaluated together. The company found Excel far too cumbersome for planning, and it gave no understanding of how these decisions affected product and customer profitability. The company also could not account for fluctuating exchange rates and numerous currencies in its decisions. The business clearly needed an overhaul of its planning and analytics approach.
The Solution
With the EO Decision Engine, the company now has all of its mining and manufacturing processes represented in a single model of the business. The model shows how the business works from mine to customer and describes the financial components of the cycle at every step. Every recommendation is grounded in that model, so planners can see which constraints and trade-offs drive it.
The company uses two planning and decision-support solutions for mid-term and long-term planning.
Monthly planning (24-month horizon). This solution shows which products are most profitable on a per-product and/or per-customer basis. It answers questions like:
- Should we sell, process, or buy what is mined? What about chemicals that are co-produced during processing?
- What gets upgraded, and what gets sent to the chemical processing plants?
- What quality Slag/SR should be produced?
- Should we process certain products near the mine, or transport the lower quality?
- Which feedstock goes to which chemical processing plant?
- Which lines, at which plants, should make which quality or grade of product?
Strategic planning (30+ year horizon). This solution supports a yearly planning process. It shows, year by year, the best way to get starting products over the next 30 years, given the geologists' existing mine plans and future market predictions. It answers questions such as:
- Do we have enough mines? What if we opened more? Where and when should they open?
- What are the benefits of upgrades versus new structures?
- Where should we spend our capital?
- How will these decisions affect timing and yields?
- Which mine should feed which plant or plants?
- Should we automate? If so, where, and what is the impact on KPIs?
The company's modeling complexity calls for true optimization together with validated financial data. The EO Decision Engine supports both. The business was modeled and validated in a pilot phase before being scaled across the enterprise.
The Impact
With the EO Decision Engine, the company can validate the best courses of action, resulting in a 3-5% improvement in Net Present Value (NPV) within the first 6 months of deployment. The engine continues to deepen the company's understanding of its business and planning process by exposing previously hidden issues in the current plan, such as:
- Forecast sales prices that cause production swings that do not make sense
- Sales price forecasts that seem unrealistic
- And more
3-5% improvement in Net Present Value within the first 6 months
About River Logic
River Logic builds the Enterprise Optimizer® (EO) Decision Engine, which turns business goals into defensible, execution-ready decisions, each with the receipt: the binding constraints behind it and a trace back to the source data. Built on 25+ years of optimization R&D, one model of the business answers questions across planning domains, from network design and capacity to product mix and capital allocation.






















