Abstract:Optimization plays an important role in chemical engineering, impacting cost-effectiveness, resource utilization, product quality, and process sustainability metrics. This chapter broadly focuses on data-driven optimization, particularly, on model-based derivative-free techniques, also known as surrogate-based optimization. The chapter introduces readers to the theory and practical considerations of various algorithms, complemented by a performance assessment across multiple dimensions, test functions, and two chemical engineering case studies: a stochastic high-dimensional reactor control study and a low-dimensional constrained stochastic reactor optimization study. This assessment sheds light on each algorithm's performance and suitability for diverse applications. Additionally, each algorithm is accompanied by background information, mathematical foundations, and algorithm descriptions. Among the discussed algorithms are Bayesian Optimization (BO), including state-of-the-art TuRBO, Constrained Optimization by Linear Approximation (COBYLA), the Ensemble Tree Model Optimization Tool (ENTMOOT) which uses decision trees as surrogates, Stable Noisy Optimization by Branch and Fit (SNOBFIT), methods that use radial basis functions such as DYCORS and SRBFStrategy, Constrained Optimization by Quadratic Approximations (COBYQA), as well as a few others recognized for their effectiveness in surrogate-based optimization. By combining theory with practice, this chapter equips readers with the knowledge to integrate surrogate-based optimization techniques into chemical engineering. The overarching aim is to highlight the advantages of surrogate-based optimization, introduce state-of-the-art algorithms, and provide guidance for successful implementation within process systems engineering.

What problem does this paper attempt to address?

### What problems does this paper attempt to solve? This paper aims to solve optimization problems in chemical engineering, especially in the fields of data - driven optimization (DDO) and model - based derivative - free optimization (MB - DFO). Specifically, the paper focuses on how to use the collected data to optimize complex systems without explicit derivative information, especially those systems involving expensive black - box functions. #### Main problem description 1. **Limitations of traditional optimization methods**: - In chemical engineering, traditional optimization methods rely on algebraic expressions or knowledge - based models and are optimized through derivative information of analytical expressions. However, with the development of digital technologies, such as intelligent measurement devices, process analysis techniques, sensor technologies, cloud platforms and the Industrial Internet of Things (IIoT), more and more optimization problems need to be carried out completely relying on the collected data. - For complex chemical systems, data collection can often only be achieved by evaluating expensive black - box functions, which may represent in - vitro chemical experiments, costly process reconfigurations, or simulations such as computational fluid dynamics (CFD) or quantum mechanics calculations. These deterministic models usually rely on complex and expensive simulations, and due to the presence of computational noise, it is difficult to calculate numerical derivatives reliably. 2. **The need for data - driven optimization**: - Many engineering optimization challenges can be framed as "expensive" black - box problems, which are limited by the number of function evaluations. When gradient information or cost - effective models are not available, efficient optimization methods that rely only on function evaluations need to be used. - Data - driven optimization (DDO) emerged in this context, which can optimize functions by exploring the space and using existing information without analytical expressions. 3. **Model - based derivative - free optimization methods**: - Model - based derivative - free optimization (MB - DFO), also known as surrogate - based optimization, is a method for optimizing functions without using explicit derivative information. It constructs a surrogate model to approximate the target function and uses this surrogate model to guide the optimization process. - This method is especially suitable for cases where the target function is unknown, the evaluation cost is high, and there is noise. #### Research objectives 1. **Introduce and explain model - based derivative - free optimization methods**: - The paper introduces the basic concepts, theoretical foundations, and practical applications of model - based derivative - free optimization. It covers a variety of popular algorithms, such as Bayesian Optimization, Constrained Optimization by Linear Approximation (COBYLA), ENTMOT (Optimization Tool for Decision Tree Surrogate Models), Stable Noisy Optimization by Branch and Fit (SNOBFIT), etc. 2. **Performance evaluation and comparison**: - The paper conducts performance evaluations on different algorithms, covering multiple dimensions, test functions, and two chemical engineering case studies: one is a high - dimensional stochastic reactor control problem, and the other is a low - dimensional constrained stochastic reactor optimization problem. These evaluations reveal the performance and applicability of each algorithm, helping readers choose the algorithm most suitable for a specific application scenario. 3. **Provide practical guidance**: - By combining theory and practice, the paper provides readers with specific guidance on applying surrogate - based optimization techniques to chemical engineering. It not only introduces the mathematical foundations and descriptions of various algorithms but also provides GitHub code repositories, enabling readers to reproduce and compare the effects of different algorithms. 4. **Emphasize advantages and future prospects**: - The paper summarizes the results of benchmark tests, emphasizes the most effective algorithms and their shortcomings. At the same time, it looks forward to future surrogate - based optimization research, pointing out potential research directions and development trends. In short, this paper is committed to providing chemical engineers and process systems engineers.

Surrogate-Based Optimization Techniques for Process Systems Engineering

A Survey on Applications of Optimization-Based Integrating Process Design and Control for Chemical Processes

A Novel Fast Dynamic Optimization Approach for Complex Multivariable Chemical Process Systems

Surrogate Equations of State for Equation Oriented Optimization of Polymerization Processes

Assuring optimality in surrogate‐based optimization: A novel theorem and its practical implementation in pressure swing adsorption optimization

Data‐driven decision‐focused surrogate modeling

A surrogate-based optimization framework for simultaneous synthesis of chemical process and heat exchanger network

Systematic Optimization Methodology for Simultaneous Process Design and Heat Exchanger Network Synthesis with Surrogate Models – Application to Distillation Column Complex

Surrogate-based optimisation of process systems to recover resources from wastewater

Surrogate-Based Simulation Optimization

Multi-level optimization strategies for large-scale nonlinear process systems

Process flowsheet optimization with surrogate and implicit formulations of a Gibbs reactor

A surrogate‐based framework for feasibility‐driven optimization of expensive simulations

Real-time Optimization for Chemical Processes Based on On-Line Modeling of Controlled Variables

Advanced Modeling and Optimization Strategies for Process Synthesis

Advances in surrogate based modeling, feasibility analysis, and optimization: A review

Development of hybrid surrogate model structures for design and optimization of CO 2 capture processes: Part I. Vacuum pressure swing adsorption in a confined space

A Brief Introduction to Chemical Reaction Optimization

Ensemble of surrogates in black-box-type engineering optimization: Recent advances and applications

Human-Algorithm Collaborative Bayesian Optimization for Engineering Systems