Research On Supply Chain Risk Early Warning Based On ACO-SVM Algorithm

In recent years,due to the rapid development of technologies,such as transportation,information,the links between enterprises become closer and faster.And the macroeconomic environment is in a state of continuous fluctuation.Those make the sources of supply chain risk enterprises face become more complicated and supply chain risks dynamic change at a faster rate.Supply chain is a system,and the deviation occurred in any node of the link will be likely to make the whole supply chain fluctuate and the enterprises lose.Therefore,it is necessary to continuously monitor,identify,judge and analyze all kinds of sources of supply chain risk,and promptly deliver the analysis results to the relevant managers,so that the managers can take appropriate measures and propose some management plans of supply chain risk to ensure continued and efficient operation of the whole supply chain,and quickly meet the consumer’s requirement.Under this background,this paper was on the basis of previous studies,Adopt methods of literature research,fuzzy reasoning and the other,and build the early warning index system of supply chain risk scientifically and rationally in view of the typical supply chain,.In the meanwhile,it establishes a risk early warning model based on ant colony algorithm and Support Vector Machine,meeting the dynamic warning needs of the supply chain risk,and bringing some reference value to the risk warning of the supply chainFirstly,we sort out the concepts and theories related to the research content of the thesis,containing concepts,features and classification of supply chain risk,and the theoretical basis of the model algorithm.Secondly,according to the typically manufacturing supply chain,we identify risks from five dimensions of risk incentives among the macro environment,supplier,core business,distributors and enterprises,and initially establish the early warning index system of supply chain risk,then streamlined by the fuzzy inference tool,in order to obtain the final index system.Next,the paper chooses the regression mode of support vector machine having a good performance in the forecast area,as the main method of risk assessment,use the improved ant colony algorithm to overcome the blindness of parameters selection of support vector machine,and established a risk early warning model based on ant colony algorithm and support vector machine.Then,according to established indicator system previously,the paper collects the corresponding data and.compares two groups,first compares the support vector machine parameter optimization effects by ordinary ant colony algorithm and improved ant colony algorithm,then compares the effect on SVM models optimized parameters by improved ant colony algorithm,cross validation and SVM model not optimized parameters to verifies the validity of the model.In the end,the paper makes the empirical research on supply chain operation,analyzes levels of the supply chain risk and reasons why Supply chain risk value fluctuates,and besides,gives the corresponding management recommendations.