A novel nonlinear time-varying grey prediction framework for green transformation of manufacturing industry: Modeling of a non-equidistant perspective

Abstract

This study proposes a multi-dimensional nonlinear time-varying grey prediction framework, NEMTDNGM(1,1), to address the challenges of modeling non-equidistant effects in the green transformation of land and marine manufacturing industries (LMMIGT). Using technological innovation as an input, NEMTDNGM(1,1) effectively identifies its non-equidistant drivers within LMMIGT. By introducing dual time-varying effects through grey action and development coefficients, the model thoroughly examines LMMIGT’s nonlinear features characterized by energy consumption. Robustness is tested with Monte Carlo simulations based on the Marine Predator Algorithm, employing β-convergence and kernel density estimation. Empirical results demonstrate that NEMTDNGM(1,1) achieves the training and test MAPEs below 3%, significantly outperforming other models, underscoring its potential to support sustainable development of land and marine manufacturing systems.