A nonlinear grey combined model for forecasting port container throughput in the post-pandemic era

Port container throughput is a barometer of a region’s economic development. Therefore, accurate throughput forecasting is strategically important for efficient port and shipping operations, international trade coordination and regional logistics resource allocation. The COVID-19 pandemic made 2020–2022 container throughput data unrepresentative. Since the World Health Organization declared the end of the pandemic in 2023, container throughput data have been characterised by short time series, significant fluctuations and nonlinearity. To fully excavate the potential information of throughput data, this study proposes a nonlinear grey combined model (NGCM) that introduces a linear weighted genetic algorithm to establish a dual-objective optimisation strategy that simultaneously considers prediction accuracy and stability. To validate the proposed model, this study uses monthly data from major Chinese ports from 2023 to 2025. It employs four basic grey models and constructs a comparison system of 11 models through a combination of linear and nonlinear methods. The results of MAPE, RMSE, RRMSE and statistical tests show that NGCM outperforms all comparative models in accuracy and stability, successfully solving the problem of port throughput prediction in the post-pandemic era.