水网地区危化品应急资源优化配置模型

Optimal allocation model for storing emergency chemical resource of hazardous nature in the water network regions

我国水网地区航道纵横交织成"网状",具有联通性、聚集性及应急救援路径选择多元性等特殊性,对水网地区危化品应急资源进行优化配置,能够对事故作出快速响应,减小事故损失,提高应急救援服务效率和水平。首先利用K-Means聚类算法对事故点进行聚类,确定应急资源点位置,完成选址模型;然后在此基础上,以最小化应急资源数量为目标,建立基于随机规划的应急资源分配模型,并利用遗传算法求解。将某水网地区作为算例进行验证,结果表明,该模型能够很好地满足实际配置需求。

The given paper is aimed to develop an optimal allocation model for the hazardous chemicals emergency goods or resources in the inland water network regions.And,for the said purpose,the given paper has mainly focused on the following two problems：one is how to choose the proper place for placing such goods while the other is the proper amount of such goods or resources to be allocated to each point of such locations.Through such an optimized allocation model,immediate response results can be made to the corresponding accidents with the necessary efficiency of the emergency rescue measures and the service level concerned to be properly brought up forward.In addition,it is also necessary to put forward necessary navigational channels in the correlated inland water network regions with the needed advantages or preconditions,as well as the criss-cross navigational channels interwoven into the ＂network＂,such as the connectivity of the waterways,the aggregation of the ships and ports,the diversity of the emergency route selections once accidents would come about.And,all of the aforementioned factors may tend to increase the complexity of the response and rescue strategies in the said water regions.To achieve the said goals,we have in this paper,firstly,adopted K-Means clustering algorithm to cluster all the points such accidents may come about,and,then,it is necessary to determine the centers of each cluster,which should be chosen as the emergency rescue resource points to construct the site model to be chosen through the repeated iteration.And,next,aiming to minimize the amount of the emergency resources,it is necessary to establish an allocation model for the emergency resources based on the stochastic programming.And,then,the model has to be worked out through a genetic algorithm to determine the emergency resource amount allocated in each point.Thus,it would be possible to verify the actual situation in-situ of the said water networks region by taking the results of the experiment as an example,which is surely to be made enough and satisfactory.From the above description,it can be seen that the effect of the cluster should be self-evident,and the algorithm we have proposed in this paper can well be made to meet the actual allocation demands.Besides,the method developed in this paper can thus be taken as a reference for the decision makers to program the emergency resource construction and optimize the resource allocation in dealing with the hazardous chemicals in the inland water comprehensive network.