MOSAICmodeling enables a number of different advanced features based on the simulation of steady-state or dynamic system. Among these are the simultaneous and sequential optimization. A number of additional features, such as parameter estimation and chance-constrained optimization will be added subsequently.
Optimization inside MOSAICmodeling is currently limited to general Mixed-Integer NonLinear Programming (MINLP) problems:
$latex \min_{u,y} f(x,u,y)$
s.t. $latex g(x,u,y) = 0$,
$latex x^L \leq x \leq x^U$,
$latex u^L \leq u \leq u^U$,
$latex y \in \mathcal{Z}^n$,
$latex y^L \leq y \leq y^U$
Inequality constraints at the moment need to be manually reformulated as equality constraints by the introduction of slack variables $latex s$:
$latex h(x,u,y) \geq 0$ $latex \Rightarrow h(x, u, y) – s = 0$ with $latex s \geq 0$
Advanced Optimization Features
Follow the links below for detailed descriptions on how to use each feature.
- Simultaneous Optimization
- Sequential Optimization