The Research And Implementation Of Logistical Demand Forecast System Based Upon L-OD

With the development of the global economic integration, the modern logistics plays a more essential role in the national economy system. Increasingly expanding the society demand for logistics propels the central and local governments and tons of enterprises to accelerate and strengthen the logistical infrastructure construction and investment. Nevertheless, the industry needs to set up a set of theory system of the modern logistics planning and the logistical demand forecast, which are employed to analyze and solve the problems about how to realize the development rules of logistics demand, and how to deal with logistics facilities layout. In light of L-OD forecast theory and other ready-made logistics planning principles, the paper develops a logistical demand forecast system based upon L-OD, and puts the system into practice and research. Then, the application proves the theory and principles are effective and reliable.At the beginning, the paper introduces the connotation and characteristics of the logistical demand, and analyzes the fundamental principles and process of the logistical demand based upon L-OD. Relying on these rules, the paper puts forwards the main and core mathematical models applied into the four stages of the logistical demand forecast—logistical generation, distribution, logistical chain and assignment. The first stage models include: Regress analysis, Elasticity factor, Growth factor, Cross-classification analysis; the second stage models include: Mean-growth factor, Detroit factor, Fratar factor, Modified WRXS Gravity model, Modified BPR Gravity model, Double Restricted Gravity model; the third stage models include: Proportion model, Probit model; the fourth stage models include: All-or-nothing assignment model, Static Multi-route model, Limited Capacity model. In addition, the paper presents some reasonable and effective methods for the parameters calibration of the vital models, solution to models and verification of forecast. Associated with the four stages of the logistics forecast and the concrete necessities of engineering projects, the paper, furthermore, takes advantage of the object-oriented analysis based on UML to research function and data requirements of the logistical demand forecast system, and establish the system Use Cases models and the Domain models, describe the complete forecast process with activity diagram, and then, proposes the system design and implement; meanwhile, a practical engineering case is demonstrated, which utilizes the system to achieve the logistical demand forecast.Finally, the paper draws attention to the Web Services technology based on XML data transferring and exchanging technology and SOAP in the logistical demand forecast application, and studies further on dynamic Web Services composition technology. At the same time, Modified Backward Chaining algorithm is proposed, which sets up a kind of linked-list mechanism to record the route of composition and facilitates WTOP convert composition plan into BPEL4WS smoothly and efficiently. The brand-new intelligent composition technology will explore and develop the novel domain for the next generation of the logistical demand forecast system.