基于动态Holonic的虚拟应急物流系统建模及应急物资耦合协同优化调度

Emergency materials scheduling belongs to multi-objective stochastic combinatorial optimization problem in the condition of variable environment and information loss. Dynamic Holonic modeling theory will be employed in this research project for virtual emergency logistic system. Hence, the coupling evolving rules in space-time between the strength of Holon network relation intensity and the efficiency of emergency materials scheduling, characteristic of system dynamic and dynamic allocation and real-time feedback mechanism of the information collected by EPC in the system will be researched respectively. The interactions between dissimilar paths will be contained in the modeling mechanism, and the emergency material scheduling and emergency vehicle path are integrated to real-time dynamic optimization. The dynamic coupling mechanism of emergency material supply and demand, the coupling mechanism of material demand and disaster dynamic evolution, measure of temporal and spatial dissimilarity are also researched respectively, and the time-space network evolution algorithm based on the importance of the vertex, the priority of the vertex and niche replication technology will be proposed. Accumulation and transfer mechanism of parameter perturbation effect on the decision scheme will be researched. By the theory of nonlinear normal mode and stochastic reliability, sensitive of parameters will be analyzed as the foundation of scheduling scheme stability, anti-interference ability and local dynamic real-time adjustment for relevant departments to make decision schemes. trend function of swallowtail catastrophe model will be constructed to analyze the critical point of mutation and stability of capital operation efficiency of emergency materials scheduling. It wiil provide the scientific basis for decision-making of emergency materials scheduling by research this project.

应急物资调度是一类在环境变化和信息不完备条件下的多目标随机组合优化问题。基于动态Holonic系统建模思想，研究Holon网络关系强度与应急物资调度效率在时空上的耦合演化规律和系统动力学特征，探讨物联网采集的信息在系统内动态分配及实时反馈机制。考虑问题的时空复杂性、交通网络属性的时间依赖性与随机性及地区特征，研究应急物资供需动态耦合机理、物资需求量与灾害动态演化的耦合机理、各运输方式共同参与应急的协同机理和时空相异性测度，将相异路径间的相互影响蕴含于建模机理，对物资调度和应急车辆径路集成实时动态优化，构建实时调度的非线性空间时序优化模型及基于顶点优先权、顶点重要度和小生境复制技术的时空网络进化算法。研究参数扰动对决策方案影响的积累和传递机理，提出分析调度方案抗干扰及方案局部动态调整或更新的理论依据和策略。构建燕尾突变模型的趋势函数，分析调度效率的突变临界点及稳定性。为应急决策提供科学依据。

应急物资调度是一类在环境变化和信息不完备条件下的资源动态调度问题，也是调度理论中的难点问题，目前仍没有科学的解决方案。课题组通过对该项目的研究，取得的研究成果包括：（1）通过引入虚拟运作理论，提出了虚拟应急物流协同体系VELCS；通过引入Holonic系统建模思想，集成递阶系统的稳定性和混杂系统的柔性，构建了基于Holonic系统的VELCS的虚拟应急物流运作体系及系统模型所依赖的动态交互机制。（2）考虑了路段剩余容量和物资分配两个因素对应急物资运输路径选择的影响，充分利用大数据时代背景下各路段和灾区先验信息的可得性，运用Bayes决策理论和UE模型建立了拥挤条件下应急物流网络优化的双层模型。（3）引入超网络理论，以应急物资运输时间最短为目标，建立了基于超网络的应急物流网络多径路优化模型，通过各层网络间的相互作用，实现网络对复杂运输环境的感知能力和反应能力。（4）考虑到突发事件具有突发性、非例行性、不确定等非常规特点，通过案例推理（CBR）以及本体（Ontology）等理论，提出了应急资源动态需求分析的相关策略。（5）考虑到突发事件条件下各类复杂不确定因素对决策方案的扰动，揭示了各类复杂不确定因素的扰动对决策方案影响的内在机理。（6）考虑到应急物流高可靠性的特征，以及构造复杂系统可靠性指标和参数间解析表达式的困难性，提出了基于贝叶斯网络理论、动态贝叶斯网络（DBN）理论、模糊动态贝叶斯网络（FBDN）的应急物流可靠性分析策略。