Syngas Production Optimization via Advanced Metaheuristic Algorithms: Simulated Annealing and Differential Evolution

The growing interest in renewable energy sources has grown over the last decade as a result of the desire to reduce harmful effects to the environment and effective resource management. Syngas, a mixture of carbon monoxide and hydrogen, plays a significant role in clean fuel production. It is widely used for electricity generation, steam production, transportation fuels and as a feedstock for production of fertilizers, chemicals and consumer products. Syngas can be produced through various processes, including steam reforming, carbon dioxide reforming, and partial oxidation of natural gas. However, these processes are affected by several variables such as feedstock composition, temperature, pressure, catalyst, feed rate, and flow velocity. Therefore, determining the optimal operating conditions for syngas production is important to improve process efficiency, reduce operating costs, and minimize time-consuming trial-and-error experimentation. In this context, the problem addressed in this study is the simultaneous optimization of multiple conflicting syngas production objectives under process-variable constraints, which remains a challenging issue in chemical process optimization. Despite the growing use of metaheuristic approaches, the application of Simulated Annealing (SA) and classical Differential Evolution (DE) to this specific multi-objective syngas production problem has not been systematically explored in the literature. In this study, SA and DE were applied to identify the optimal operating parameters for syngas production. The decision variables were oxygen-to-methane ratio, gas hourly space velocity and reaction temperature, while the objective functions were methane conversion, syngas yield, and the H2/CO ratio. Algorithm performance was assessed using a normalized Euclidean distance metric to the ideal point. This study presents the first systematic application of SA and DE to multi-objective syngas production optimization, introduces a normalized distance metric for balanced multi-objective evaluation, and provides benchmarking across five metaheuristic algorithms. The results show that both algorithms substantially outperform existing methods in the literature, with DE and SA achieving Euclidean distances of 2.6855 and 2.6916 to the ideal point, respectively, representing improvements of 35.7-56.6% over previously reported algorithms. DE provided the best overall performance, while SA produced a solution only about 0.2% inferior to DE, despite its simpler single-solution search structure. These findings indicate that the proposed approaches are effective for this benchmark syngas optimization problem and also provide methodological insight into the relative performance of different metaheuristic search strategies in chemical engineering applications.