System dynamics modelling for improving regional logistics integration: A case study of western China

Regional logistics integration (RLI) has emerged as a pivotal driver of regional integration (RI), playing a critical role in fostering regional coordinated development. However, research on RLI operational mechanism has not been systematically explored, which limits the proper assessment of RLI level under various policy scenarios, thereby hindering the effective implementation of relevant policies. To address this gap, this study analyzes empirical data from western China through a dual-validation framework, employing system dynamics (SD) modeling for scenario simulation and utilizing the gravity model alongside historical data for validation, thereby enabling systematic examination of RLI dynamic evolution under diverse policy scenarios. First, the RLI level is assessed through a comprehensive indicator system and gravity model, which serves for dual validation purposes in the subsequent SD modeling. Second, a system framework for RLI is developed based on core-periphery theory to elucidate the causal relationships among related variables. Then, a SD model is constructed and optimized to simulate RLI changes in western China up to 2035. Finally, both single-policy and combined-policy scenarios are examined, with RLI in western China being enhanced through adjustments to endogenous variables. The results indicate that the impact of single logistics soft policies on RLI becomes more significant in the later stages of the study, while the benefits of single logistics hard policies are more pronounced in the earlier stages. However, combined policies produce effects that diverge from a mere linear aggregation of single policies impacts. Notably, the systematic integration of the three types of policies is most conducive to the long-term development of RLI. These findings provide valuable insights for policymakers aiming to improve RLI. The proposed RLI model incorporates rich information, enabling policymakers to adjust the model parameters to reflect changes in complex environments, thereby facilitating the formulation of optimal RLI policies.