Advanced Numerical Methods for Stochastic Modeling in Industrial Systems
Journal of Computational Mathematics, Vol. 42, pp. 445–462
Abstract
We develop advanced numerical methods for stochastic modeling of industrial fluid and particle systems, with emphasis on stability, scalability, and uncertainty quantification under Ghanaian and West African operating conditions.
The framework couples stochastic differential equations with adaptive solvers and data-informed closures. Benchmarks on epidemic time series, climate-informed flows, and industrial cooling problems demonstrate accuracy gains over classical Monte Carlo baselines while remaining computationally tractable for institute-scale research workflows.
Results highlight practical guidance for NIMS researchers deploying stochastic models in SCIM and DSCHANGE collaborations, including reproducible solver settings and open validation datasets.
Introduction
Stochastic processes have long served as the mathematical backbone for understanding uncertainty in engineering and industrial systems. In the context of Ghana's industrial growth, the need for robust, scalable modeling tools has never been more critical. Traditional methods often fail to account for the intermittent nature of power supply and variable input quality characteristic of developing markets.
This paper establishes a framework that treats these environmental variables not as noise, but as fundamental parameters within the stochastic differential equations (SDEs) governing the system. We begin by reviewing the limitations of classical Runge–Kutta methods in high-variance environments, then introduce adaptive step-size controls and Bayesian closures tailored to NIMS research workflows.
Read the complete research
Full text available via NIMS open access links.
