Study On Key Technologies And Cloud Platform Implementation Of Agricultural Machinery Dynamic Allocation For Cooperatives

Due to the promotion of agricultural machinery and land circulation,agricultural machinery cooperatives have gradually become the main mode of agricultural production in China.However,agricultural machinery cooperatives still have some problems such as unreasonable allocation of agricultural machinery resources,backward allocation methods and lack of management platform.In order to meet the actual demand,the deep loosening machine was taken as the research object,the agricultural machinery allocation model was constructed,the agricultural machinery dynamic allocation method and data pretreatment method were proposed,and the intelligent allocation cloud platform was developed.The main research contents are as follows:(1)The dynamic multi-objective allocation model of deep loosening machine with was established by considering the attributes of farmland and farm machinery and the space-time information in the actual operation scene.The model includes transfer cost,time window penalty cost,working cost and cost based on supply and demand forecast.(2)The NSGA-Ⅲ algorithm was improved,and the delay centralized allocation strategy was proposed,which was applied to the solution of the dynamic allocation model of deep loosening machine.Based on the real operation data,a comparative test was designed to verify the effectiveness of the allocation model and the solution algorithm.The results show that the improved algorithm is effective and has better comprehensive index than the previous algorithm.(3)Based on the sliding window mechanism,an online data repair method was proposed,which was in line with the spatio-temporal correlation characteristics of agricultural machinery operation data.The validity of abnormal data identification and data repair is verified by test,and the algorithm parameters are optimized.The test results show that,when the data scale reaches 1×105 and the proportion of abnormal data is 5%,the anomaly recognition rate of the algorithm reaches 0.94.The optimized algorithm can achieve the root mean square error of 0.16 and the running time of 0.13s respectively,which can meet the requirements of the platform.(4)In light of design principles of manual interaction software,the functional and non-functional requirements of the platform were analyzed,the dynamic allocation algorithm of agricultural machinery and the pretreatment algorithm of agricultural machinery operation data were integrated,and the intelligent allocation cloud platform for agricultural machinery was developed.The function and performance tests were carried out to verify the practicability and reliability of the cloud platform.In general,a dynamic multi-objective allocation model of deep loosening machine is established,a solution method is proposed,and intelligent allocation cloud platform is completed,which can help solve the problems of lagging in information release,unreasonable allocation of agricultural machinery,and low operating efficiency in the process of agricultural machinery operations,and can provide scientific and effective decision-making solutions for agricultural machinery cooperatives.