Interactive Multi‐objective Dynamic Optimization of Bioreactors under Parametric Uncertainty

Abstract

Model‐based optimization techniques play a key role in achieving a sustainable operation of biochemical processes. Models are an approximation of the real process under study, hence, uncertainty is inherently present and for a sustainable process operation this uncertainty should be accounted for. In practice, optimality with respect to different conflicting objectives is required and multi‐objective optimization is a valuable tool. In this article the sigma point approach is applied to account for parametric uncertainty in the frame of interactive multi‐objective bioprocess optimization.

Toward Performance Improvement of a Baculovirus–Insect Cell System under Uncertain Environment: A Robust Multiobjective Dynamic Optimization Approach for Semibatch Suspension Culture

“…23 These uncertainties can arise due to several factors such as parameter estimation using noisy data, external process disturbances, high degree of nonlinearity, limited experimental data causing model identifiability problems, model assumptions, etc. 23 Thus, incorporation of such uncertainties is important while performing the optimization in order to obtain robust control profiles (i.e., manipulated variables during the process), which guarantee better objective estimates without violation of any constraints. In the literature, a number of approaches for handling uncertainty while performing optimization have been reported.…”

“…Although such a modeling and optimization exercise can offer significant advantages for large-scale process development, in practice, this analysis can lead to suboptimal or even infeasible estimates without considering the uncertainties which are inherently present in the model . These uncertainties can arise due to several factors such as parameter estimation using noisy data, external process disturbances, high degree of nonlinearity, limited experimental data causing model identifiability problems, model assumptions, etc .…”

“…Optimization under uncertainty has also been applied for other biological processes. ,− Liu et al applied the concepts of parametric uncertainty for ensemble modeling of a hybridoma cell culture for maximizing monoclonal antibody production . Optimization under uncertainty has also been performed for various biochemical networks with single objectives as well as the conflicting objectives of minimizing the energy consumption while maximizing the production of a specific metabolite. , Logist et al applied multiobjective dynamic optimization for the optimal design and operation of a jacketed tubular reactor and a fed-batch bioreactor with conflicting objectives .…”

Analysis of Weighting and Selection Methods for Pareto-Optimal Solutions of Multiobjective Optimization in Chemical Engineering Applications

“…MOO has attracted increasing number of applications in chemical engineering. 2 Recent papers on MOO applications by active researchers in process systems engineering are design and control of intensified distillation sequence, 4 model simulation of ethylene oxide formation in a packed bed membrane reactor, 5 energy system design for a commercial building, 6 resilient design and operation of process systems, 7 generation of rule-based classifier for categorical data, 8 design of solid oxide fuel cell with a gas turbine hybrid system, 9 design of biomass supply chains, 10 interactive MOO of biochemical processes with parametric uncertainty, 11 design of carbonhydrogen-oxygen symbiosis networks, 12 sustainable water management, 13 and batch distillation optimization for economic and environmental objectives. 14 Our group has been active in MOO research for about 2 decades and has contributed two books 15,16 and many articles on MOO and its applications.…”

Making sense of parameter estimation and model simulation in bioprocesses

“…Also, in recent years, some effort has been focused on accounting for model uncertainty during bioprocess optimization and control, as the cases of Liu and Gunawan (2017), Bradford, Schweidtmann, Zhang, Jing, and del Rio-Chanona (2018), Pantano, Fernández, Serrano, Ortiz, and Scaglia (2018), Nimmegeers, Vallerio, Telen, Van Impe, and Logist (2019) and Neba, Tornyeviadzi, Østerhus, and Seidu (2019). All these new developments show that quantifying the uncertainty of a kinetic model is key on its applicability for process design.…”