APROS-NIMBUS: Dynamic Process Simulator and Interactive Multiobjective Optimization in Plant Automation

Coupling dynamic simulation and interactive multiobjective optimization for complex problems: An APROS-NIMBUS case study

“…When using the interactive method, the DM was forced to consider the trade-offs when providing his preferences in step 4 of the AIMO algorithm and when the obtained Pareto optimal solutions did not follow his expectations, the question of problem model validity was brought up. The author has had similar experiences with other studies, but they have not been reported beyond [PIV,89] and this thesis.…”

“…For example, as reported in [PIV], an example of successful validation of the problem model by applying the AIMO algorithm can be found in [89], but the AIMO algorithm is not mentioned there. In [89], mistakes in the problem model formulation were detected immediately when the AIMO algorithm was applied, yet regardless of that, the problem model was previously solved with several different a posteriori methods. When using the posteriori method, the set of Pareto optimal solutions was shown to the DM, but he was not provided with any support in choosing the most preferred solution from the set.…”

“…The NIMBUS method has been used for solving a wide array of multiobjective optimization problems, including chemical process design [32], optimal shape design of ultrasonic transducers [39], separation of glucose and fructose [33], continuous casting of steel [57,58,65] intensity modulated radiotherapy treatment planning [79], brachytherapy for cancer treatment [81], optimization of heat exchanger network [44], optimizing configurations of an oxyfuel power plant process [100], wastewater treatment plant simulation and optimization [35,36], dynamic optimization of a two-stage separation process [PIV] and for concentration control problem [89].…”

“…Instead, problems modeled with different simulators or programming languages can be connected to the framework with different integration tools. Examples of the use of such integration tools include the MAT-LAB® environment used for radiotherapy treatment planning [79,80], the Balas® process simulation software for chemical process design [32,34], the Numerrin environment used in optimal shape design [39], the GAMS environment used for optimization of a heat exchanger network [PI,44] and an oxyfuel power plant process [100], the GPS-X™ simulator used for the optimization of operating a wastewater treatment plant [PIII, 35,36], the APROS dynamic process simulator used for optimization of a two-stage separation process [PIV] and for plant automation [89] and almost any programming language, such as C, C++, and Fortran [40,49,58,65].…”

“…These have included a wide spread of different types of optimization problems, including the optimal design of the structure of a grapple loader [7], decision support for a housing mobility program [41], problems in chemical process design [32], the separation of glucose and fructose [33], the optimal control of a continuous casting of steel [57,58], in-tensity modulated radiotherapy treatment planning [79], brachytherapy for cancer treatment [81], the optimization of a heat exchanger network [PI, PII,44], optimizing configurations of an oxyfuel power plant process [100], optimizing the concentration-dampening process in a tank [89], the optimization of an ultrasonic transducer [39], paper making [50], operating a wastewater treatment plant [PIII, 35,36], the optimization of an internal combustion engine [4], optimal design and control of a paper mill [93], the design of a permanent magnet direct-driven wind generator [29], and the optimization of a two-stage separation process [PIV, PV], among others.…”