Forecast Analysis of Import and Export Economic Trade based on IPM Real-Number Coded Genetic Algorithm

This study addresses the challenges of forecasting import and export trade amid global economic complexity by proposing an Improved Paper Model (IPM) real-number coded genetic algorithm. Traditional models like ARIMA and LSTM often struggle with nonlinear, time-varying economic variables and multivariate interdependencies. The IPM algorithm enhances genetic operations through real-number coding, adaptive fitness function optimization, and improved crossover/mutation strategies. Experiments using China’s decade-long trade data demonstrate the model’s superiority: it achieves lower RMSE (35.62 vs. 38.14/42.67) and MAE (28.47 vs. 31.52/36.15) and higher R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> (0.92 vs. 0.89/0.85) compared to LSTM and ARIMA. Stability tests reveal smaller standard deviations in errors, and multivariate analyses confirm robustness under GDP, exchange rate, and tariff fluctuations. The IPM algorithm offers higher accuracy, adaptability, and stability for complex trade forecasting, though scalability and dataset diversity require further exploration. This work provides a reliable tool for policymakers and businesses navigating volatile global markets.