货运量是制定物流规划、物流政策和确定物流基础设施及投资规模的主要依据。货运量预测的可靠性将直接影响到货运政策的科学性和物流基础设施投资的收益情况。目前用于货运量预测的模型主要有趋势及时间序列预测模型、系统动力学模型、交通率模型和投入产出模型等，这些模型有其应用优势，也有其理论缺陷，例如没有考虑经济关联、货运交通成本、政策影响和供需及物流行为，并且无法得到产生吸引量，与客运模型交互能力差等。其中最主要的一个缺陷就是没有考虑货运交通成本和与货运交通成本相关的货运政策的影响。这个缺陷使得现有理论无法考虑货运交通成本降低带来的诱增货运量，因此现有理论的预测值与实际相比普遍偏低。本文通过对目前的货运量预测模型和理论进行回顾，并总结各类模型的研究和应用现状，通过对比分析及其他学者的研究得出其优劣，在反思其缺陷的基础上，提出了基于可计算一般均衡的货运量预测模型。可计算一般均衡模型本是计量经济学中的理论成果，常用于评估财政税收等政策影响。本文通过对可计算一般均衡理论的应用，建立了生产函数、贸易函数以及经济主体行为函数，通过一般均衡关系进行货运量预测。这个模型考虑了内生商品价格，在内生商品价格中加入了货运成本从价率，使得模型能够考虑货运交通成本的货运量预测模型，同时能预测产生吸引量，可有效解决当前预测模型的缺陷。模型建立之后，以重庆为案例进行了实证分析，通过数据收集、SAM表编制、参数确定、GAMS程序编写等步骤进行重庆市货运量的预测。与实际货运量比较和其他文献预测值比较可以看到模型的预测效果良好。通过与其他文献的对比，发现基于可计算一般均衡的货运量预测模型相对传统模型确实能够预测货运政策诱增量。本研究的主要贡献在于提出了基于可计算一般均衡的货运量预测模型，解决了大部分现有预测理论的不足，创新了货运量预测理论，并有效探索了经济学和货运物流的学科融合。

Freight volume is the main indicator in freight planning, policy-making and investment of freight-related infrastructures. Current freight forecasting models including trend and time series, system dynamics, input-output model and zonal trip rate model, have their disadvantages as well as advantages. Little insight into freight-economic linkages and limited scope for freight cost related policy lead to the exclusion of induced freight by freight cost reduction, making estimation from current models inconsistent with the facts.This study reviews the state of the art of current freight forecasting models by referring to other scholars’ researches, and then proposes the CGE-based freight forecasting model. Computable General Equilibrium (CGE) is a theory from econometrics and usually used in the appraisal of impact of policy in finance and tax. This study applies CGE by setting up production function, trade function, behavior function that describes economic entities, the freight volume as well as production/attraction volume are obtained when equilibrium is reached. The model considers commodity price as endogenous and an ad valorem is incorporated to embody freight cost, thus the disadvantages of current models are improved by the proposed model.The demonstration is about Chongqing City. The model output a very good result comparing to actual values and values from other scholars. We can find that induced freight by impacts of freight-cost related policy is indeed reflected in the model, thus current forecasting models are improved.The main contribution of this study lies in improvement of the disadvantages of current models using proposed CGE-based freight forecasting model and the probes into the multi-subjects amalgamation.