pyMechOpt: A Python toolbox for optimizing of reaction mechanisms

Abstract

This paper introduces pyMechOpt, an open-source Python package designed for the optimization of chemical reaction mechanisms. The package implements a range of optimization methods, including conventional algorithms such as genetic algorithms (GA) and particle swarm optimization (PSO), as well as introducing novel methods such as coordinate descent (CD) and multi-objective optimization algorithms. The optimization of a reduced GRI-Mech 3.0 mechanism for methane combustion is used to demonstrate the capabilities of pyMechOpt. The SILSCD method demonstrated a notable reduction in the objective functions, exceeding the capabilities of other methods. In the context of multi-objective optimization, NSGA-III demonstrated a balanced Pareto front, outperforming both CTAEA and MOEAD. These results serve to illustrate the efficacy of the novel methods implemented in pyMechOpt. The package provides a versatile platform for researchers to customize optimization algorithms and objective functions, supporting detailed analysis of results. This package makes a contribution to the field by introducing innovative optimization methods and a comprehensive software tool for refining chemical reaction mechanisms.