Research On The Processing And Prevention Of Exceptional Events In Warehousing

In the operational management of warehousing,many unpredictable exceptional events occur frequently,which often lead to the interruption of warehousing,increase of storage logistics costs and immense waste of time and human resources,especially in a developing warehousing system.Exceptional events in warehousing are practical production problems that are similar to emergency events in supply chain.However,most existing research on warehousing is mainly concentrated on normal activities,while pays little attention to exceptional events,which is limited to exceptional monitoring and early warning.In order to perfect the warehousing management,a comprehensive idea from before and after the exceptional events in warehousing is presented.Therein,clustering analysis is adopted in the processing after the exceptional events,but existing algorithms cannot satisfy the three conditions simultaneously that are required for the classification of exceptional events in warehousing.Therefore,in this paper,a systematic and complete research idea for exceptional events in warehousing is provided,and a suitable clustering algorithm for the processing after the exceptional events is also provided.The main work is as follows:(1)For the exceptional events in warehousing,the idea for classification processing after the events and effective prevention before the events is proposed.Based on the related researches such as emergency events in supply chain,the concept definition,feature description and preliminary classification of exceptional events are summarized,which constitute the research foundation.For the classification processing after the events,a framework of classification flow based on clustering analysis is constructed.In this framework,events classification is completed by using C# language in.NET platform,and the final classification result is determined by using statistical rules.For the cause analysis before the events,a framework of cause analysis flow based on the classification result is constructed.In this framework,the cause index system based on Accident Cause Theory is established,the key causes are selected by utilizing Fuzzy Comprehensive Evaluation based on AHP,and the causal relationship and hierarchical structure are analyzed by utilizing the DEMATEL-ISM method.The above two frameworks are applied to case study,and the specific solving strategies for exceptional events in warehousing are presented after analysis,which illustrates the feasibility of research idea.(2)For the classification processing after the exceptional events in warehousing,an integrated clustering analysis method which is suitable for practical production problems is proposed.In order to satisfy the three conditions for the classification of exceptional events,the unified similarity metric for mixed-valued attributes dataset,the non-randomization determination approach for initial clustering centers and the automatic determination mechanism for number of clusters are integrated,and the penalized competitive clustering algorithm based on the standardized initial clustering centers(SIC-PCC)algorithm is proposed.The performance comparisons of SIC-PCC algorithm and two existing algorithms are conducted from algorithm characteristics,difference rate and number of clusters and clustering accuracy,which verify the superiority and applicability to the practical problems of the SIC-PCC algorithm.