购物出行与交通方式组合出行链仿真分析

居民交通出行活动研究是制定城市交通政策和规划的重要基础。考虑了出行活动一系列的选择过程,将出行活动和交通方式按照发生序列创建组合链,组合链包含了出行目的、交通方式、活动序列等信息。通过组合链可以得到购物出行频率和交通方式指标。以交通出行需求条件弹性较大的购物出行为基础,利用居民一日出行活动日记数据,提取购物出行组合链。利用神经网络并行处理及非线性拟合的优势构建组合出行链预测模型,分别建立了基于传统神经网络、LM神经网络和切比雪夫神经网络的仿真模型,输入出行者社会经济属性、家庭特征等17个影响因素对组合链模型进行训练,经验证切比雪夫模型的预测效果最优。标定后的模型可以较好的预测购物出行频率和交通方式选择,可为交通需求管理政策研究、效果预判等提供量化分析工具。

Development of a Portfolio Simulation System for Apparel Supply Chain

In this paper,a generic retailer-oriented portfolio simulation system linking manufacturer,retailer,and customer in apparel supply chain is presented.The purpose of the simulator is to generate a portfolio consisting of replenishment strategy and performance index under different sales forecasting errors to satisfy the retailer-defined customer service level in apparel industry.After analyzing the main parameters in the portfolio simulator,the procedure and detailed structure of the simulator are then described.With the use of data from the industry,one case study of the portfolio simulator is achieved and the process of the simulation is validated.