On this page we present the software components of the PAROC framework. In general, any high-fidelity action is performed in PSE's gPROMS ModelBuilder, while all other tasks are executed in MATLAB ®. Our software solutions are tailored to work with these two programs and to allow for the application of PAROC. The key components are thereby:
Based on the approximate model, formulate the corresponding multi-parametric programming problem
Solve the multi-parametric programming problem
Make the solution of the multi-parametric programming problem available to gPROMS®.
Each of these tasks can be treated separately, and especially in the case of the solution of the multi-parametric programming problem, yields a stand-alone software for the solution of multi-parametric programming problems. Note: the versions which are available here will always be the ones which are up to date. If you are interested in an older version of the software, please contact us.
The POP Toolbox
The core of PAROC is its ability to solve multi-parametric programming problems. To this end, we have developed POP, the Parametric Optimization toolbox which allows for the efficient solution of multi-parametric programming problems. It features:
Algorithms for multi-parametric linear and quadratic programming (mp-LP/mp-QP) problems, namely:
Handling the integer variables using global optimization or exhaustive enumeration
Additionally to its solver capabilities, it also features a problem library as well as a problem generator, which allows for its use beyond PAROC and as a general toolbox for multi-parametric programming problems.
POP comes fully equipped with a suite of test set problems, which allow for benchmarking of algorithms and as an idea of the type and scale of problem, multi-parametric programming is currently able to solve. The latest version of POP, together with the test sets 'POP_mpLP1', 'POP_mpQP1', 'POP_mpMILP1' and 'POP_mpMIQP1' can be found here:
In order to facilitate the use of the POP toolbox, below you can find several tutorial videos related to the use of the POP toolbox:
The Controller Design
Closing the loop - mpMPC.dll
Once the solution of the multi-parametric programming problem has been obtained, its solution needs to be passed on to gPROMS, where the obtained controller is validated against a high-fidelity model. Currently, this procedure is performed via gO:MATLAB, the in-build link of gPROMS with MATLAB ®.