Defining a Stochastic Optimization Model
The following topics teach you how to model stochastic optimization models using the RASON Modeling Language.
This guide will give you step-by-step instructions on how to create and solve a stochastic optimization model but if you'd like to open the example and follow along, you can do so by browsing to C:\Program Files\Frontline Systems\Solver SDK Platform\Examples\Rason and opening the file, UGProjectSelect0.json , if using the Desktop IDE or, if using the Web IDE, click RASON Examples on the ribbon down select User Guide -- UGProjectSelect0.json.
We use the term stochastic optimization to mean optimization of models that include uncertainty, using any solution method. Solver SDK Platform and the RASON Modeling Language offer an exceptional level of power to find robust optimal solutions to models with uncertainty, using three different solution methods:
- Simulation optimization
- Stochastic programming
- Robust optimization
The first method, simulation optimization, uses the Evolutionary engine to handle very general models, but it is not scalable to large models (with thousands of variables and constraints), and it doesn't support the important modeling concept of recourse decisions. We used this method in the section above to solve the UGYieldManagement3.json model.
Stochastic programming and robust optimization can be applied only to linear and quadratic programming models with uncertainty, but they are scalable to large models.
|