The Research And Application Of Association Rule In AGV System Based On Genetic Algorithms

AGV (Automated Guided Vehicle, AGV), which is the key equipment of modernlogistics system, has attracted much attention. More and more medium-sizedenterprises start to procure AGV system and deploy the system.They try to automaticfreight robot to replace the manual work and save the cost of human resource,gradually realizing industrial automation. A lot of irregular data is accumulated duringAGV system operation. It is a problem worthy of study that how to use data miningtechniques to do effective analysis to the data of AGV systems for extract usefulinformation, and use this information to improve the operational efficiency of theAGV system. This paper brings forward a mining method of association rules basedon Genetic Algorithm,and analyses the data of the AGV system.The main work isdescribed as follows:This paper describes the relevant knowledge: data mining, genetic algorithms andassociation rules. Studying the limitation in model based on support and confidencemeasure, the comprehensibility measure and the interestingness measure are given inthis dissertation. According to the support, confidence, comprehensibility andinterestingness in degree comprehensive evaluation of the association rules.For association rules mining, the calculation problems of all the evaluationcriteria to repeat scanning database, this paper proposes a attribute catalogue.According to it, we can effectively reduce the number of scanning database andreduce the time of the association rule mining.Genetic Algorithm has the global searching capability. On this basis, this paperbrings forward a mining method of association rules based on Genetic Algorithm,meanwhile discusses and analyses the genetic algorithm in detail from coding method,the fitness function consists of support, confidence, comprehensibility andinterestingness, getting its value with attribute catalogue, initial population, geneticoperators and other aspects.Finally we apply the algorithm into analyzing the AGV system, and find somevaluable association rules. Compared with other algorithm, the algorithm proved to beeffective. We can optimize AGV system and give some valuable advice in warehousemanagement according to the rules.