Scope
- Stemly worked closely to successfully optimize cocoa blending for a plant in South East Asia
- Intricate process mapping to explicate and systematize organically developed production and decision processes
- Generate multiple scenarios by varying parameters for comparison purposes
- Assess progress against goals and refine planning – a feedback loop
Solution
- The solution contains a set of decision tools spanning multiple time horizons from the operational to strategic level:
- Annual blending plan
- Monthly blending plan
- Daily blending plan
- Supports local variation in production processes e.g.: constraints
- Accommodates variation in blending optimization goals by plant/region
- Modularity allows rapid development of additional decision tools on request
Results
- Data-driven rather than assumption-driven
- Frees up production planners to focus on planning over tedious calculation
- Quickly compare scenarios in simulation without the cost of physical experiment
- Network flow models to translate into mathematical form for a solution
- Knowledge codified to provide requirements suitable for AI-based applications
- Translates local concepts into shared production blending ‘language’.
- Distills heterogeneous data sources into a common data model
- Leverages external data for deeper context