The Combination Forecasting Model Of Logistics Demand Based On IOWHA Operator

With the economic globalization and integration accelerating, logistics as the new serviceindustry that has broad prospect has been developing throughout the world, and has more andmore influence on economy. Developing the modern logistics has obviously been anirreversible trend for the rapid improvement of national and regional economy. To build theefficiently operational logistics system and make the reasonable and effective logisticsdevelopment policies to adapt to the economic development, the logistics demand forecastingshould be done. In this context, analyzing effects of social economic activities on logisticsdemand, establishing appropriate quantitative forecast models and predicting scientificallylogistics demand can provide important basis for logistics planning and logistics situationanalysis. This paper aims at combining with the related theory and research status of logisticsdemand forecasting to establish the index system of logistics demand forecasting and selectsuitable single forecast methods. The combination forecast model is built based on singleforecast methods, and the instance forecasting and analysis are done respectively to find outthe way to improve the prediction accuracy.Firstly, the basic theory of logistics demand is summarized, including the definition,characteristics and main economic impact factors of logistics demand. Combining the currentsituation of lacking historical statistical data of logistics in our country based on consideringthe logistics demand forecasting index selection principle, the rationality of choosingeconomic indicators, freight volume and freight turnover volume as quantitative indicatorsneeded for logistics demand forecast is expounded, and the index system of logistics demandforecasting is built accordingly. Then, this paper chooses the grey forecast method and RBFneural network to construct the forecasting model according to the analysis of characteristicsof logistics demand forecasting and forecast methods. Based on this, for the problem of thelimited information and the large errors existing in the single forecasting model, and of theinvariable weights and single goal principle existing in the traditional combination forecast method, three new weights determination methods of the combination forecast model areproposed based on combining the degree of grey incidence, the vectorial angle cosine and thecorrelation coefficient and induced ordered weighted harmonic averaging (IOWHA) operatorrespectively, and the optimal weighted theory. Using the new weights determination methods,three optimal combination forecasting models are proposed based on the grey model（GM）and RBF neural network model. Finally, the logistics demand forecasting of Beijing is givenby using the single forecasting models and combination forecasting models built aboverespectively. Through comparing and analyzing, the result shows that the combinationforecasting models based on the correlation index and the IOWHA operator can improve theprediction accuracy effectively.