CAJ | 학술논문

鲜活农产品因其易腐烂、难保存的特点导致运输过程中损耗严重，成本重复，因此对运输有较高要求。通过层次分析法确定农产品运输通道评价指标权重，将变尺度混沌优化与BP人工神经网络有机结合，可构建变尺度混沌-BP神经网络组合优化模型。该模型通过变尺度混沌优化方法对网络权值分布进行优化，采用梯度下降法最终搜索确定全局最优权值。该模型可实现对农产品运输通道模式的评价和选择，避免了各种运输方式成本的重复利用，可快速有效提高农产品运输的运输质量。该模型可做到全局优化，避免了以往算法局部最优化的缺陷。通过实例仿真，该模型具有较好的应用性和动态适应性。
선활농산품인기역부란、난보존적특점도치운수과정중손모엄중，성본중복，인차대운수유교고요구。통과층차분석법학정농산품운수통도평개지표권중，장변척도혼돈우화여BP인공신경망락유궤결합，가구건변척도혼돈-BP신경망락조합우화모형。해모형통과변척도혼돈우화방법대망락권치분포진행우화，채용제도하강법최종수색학정전국최우권치。해모형가실현대농산품운수통도모식적평개화선택，피면료각충운수방식성본적중복이용，가쾌속유효제고농산품운수적운수질량。해모형가주도전국우화，피면료이왕산법국부최우화적결함。통과실례방진，해모형구유교호적응용성화동태괄응성。
Based on the combination of the variable scale chaos optimization and BP artificial neural net-work, an optimized Variable Chaos -BP neural network model ( VCBP) has been proposed. The VCBP model optimizes network weight by the method of variable scale chaos optimization, and reaches the glob-al optimum by the adoption of gradient descent method. This optimized algorithm can achieve global opti-mization, and avoid the simple BP algorithm defects caused by local optimization. Through the example simulation, the model can undertake comprehensive evaluation and selection of transportation corridor of agricultural products. Simulation shows that VCBP model has good applicability and dynamic adaptabili-ty.