The Application Of The ABC-based GA-NN Model In Aircraft Customer Service Cost Prediction

With the sustained development of China’s economy, the China’s civil aviation industry hasdeveloped rapidly. The Aircraft Customer Service Cost Prediction has played an important role inChina’s civil aviation industry development. How to predict the Aircraft Customer Service costaccurately, how to improve the Aircraft Customer Service quality and efficiently, how to reduce theAircraft Customer Service Cost are the problems that must be solved urgently in civil aviationindustry. This paper introduces the Activity Based Costing into the Aircraft Customer Service CostAccounting, proposes an ABC-based GA-NN model, which can be used to predict the AircraftCustomer Service cost. The main content of this paper is as follows:(1) Proposes a cost control program applied to the Aircraft Customer Service Structure. Thisprogram is a multi-level, cost-allocation-view-oriented and process-analysis-oriented method, whichcombines the Activity Based Costing and Activity Based Management.(2) Proposes the ABC model applied to the Aircraft Customer Service. This model bases on theoriginal function of cost accounting, takes into account the job management, business processoptimization, calculation accuracy and other aspects, at the time assigns related cost to the service costto make it more accurately, while all business processes of the service are completely described. TheABC model not only presents the trend in the business process of the Aircraft Customer Service, butalso provides the excellent training sample for prediction model. It is the key technology to guaranteethe successful of prediction model.(3) Proposes the GA-NN model applied to the Aircraft Customer Service. This paper combines thecharacteristic of genetic algorithm and BP algorithm, and focuses on the process of genetic algorithmoptimizing the Neural Network structure, the initial weights and thresholds globally, then trains theNeural Network, and uses the bidirectional method changing the hidden layer nodes to optimize theNeural Network structure constantly, so we can get an optimal model which met the convergencecondition of the target function. Finally, we use this model to predict the Aircraft Customer Servicecost.(4) Through the predictive analysis of the ABC-based GA-NN model and the Neural Networkmodel, the results prove that the prediction of the ABC-based GA-NN model is effective. Theexperiment analysis proves that the ABC-based GA-NN model can enable genetic algorithm play itsstrong search ability, and this model is more excellent than the Neural Network model in prediction accuracy and convergence speed.Finally, this paper applies the ABC-based GA-NN model in the Aircraft Customer ServiceManagement System. The application results show that this model accounts and predicts the AircraftCustomer Service Cost accurately.