With the development of the social economy, logistics demand tends to be more complicated featured by the network in the form and diversity in content, mainly reflecting in logistics integration in the network of supply chain. It can be classified as fixed and stochastic demand and can also be categorized as static and dynamic demand. The increasing complexity in logistics demand issues a new challenge to the supply system. Meeting the logistics demands in form and content, a matched logistics supply network should be built, which can response to and adapt to the changes of logistics demand.Combining the essential features of the logistics, the paper mainly studies the characteristics and properties of the supply and demand network in streamline supernetwork, based on supernetwork theory and thoughts of matching degree. It establishes the optimization models of streamline supernetwork of multi-commodity and multi-objective fixed demand, building the optimization model of streamline supernetwork of random supply and demand and the model of multi-cycle demand. By optimizing the streamline supernetwork with various features of demand, the problem of the supply matching the demand has been solved with more efficient logistics service and higher service level, meeting the increasingly complex logistics demand.For a start, the paper studies streamline supernetwork theory. It analyses the component factors of demand and supply network, researching for the types of the streamline supernetwork based on the characteristics of demand and supply network. Furthermore, the definitions of the classic variational inequalities and stochastic variational inequalities are introduced and the direct conversion relationships between variational inequalities and other problems are analyzed, while several projection algorithms are studied and stochastic projection algorithms is designed with proving its global convergence.Moreover, the paper studies the optimization model of streamline supernetwork of multi-commodity fixed demand. On the basis of fixed demand, it studies the structure of the streamline supernetwork of multi-commodity circulation and establishes the continuous cost function of the logistics demand and supply, reflecting the logistics cost of different commodities in different phrases. The optimization model, aimed at maximizing the matching degree of multi-commodity cost of supply and demand, is converted into a variational inequality problem. They are proved to be equivalence and the solution to be uniqueness. Next in the case of the streamline supernetwork of coal, a modified projection algorithm is adopted and the fact is revealed that the capacity of logistics infrastructure and the influences of the relationship among commodities on the matching degree of supply and demand.Then the paper studies the optimization model of streamline supernetwork of multi-objective fixed demand. Considering the equal importance of cost and time in logistics service, the continuous functions of the logistics cost and time cost are built based on fixed demand. The double objective optimization model, aiming to maximize the matching degree of the cost of supply and demand and the time cost of supply and demand, is converted into a related variational inequality problem. They are proved to be equivalence and the solution to be uniqueness. Next taking the streamline supernetwork of distribution between cities and rurals as an example, a modified projection algorithm is adopted and the conclusion is drawn that the key factors influencing matching degree of supply and demand are the capacity of the logistics supply system and the waiting and retention of supply time.Next the paper studies the optimization model of streamline network of random supply and demand. The logistics network of random demand is established and analyzed on the basis of the end customers’random demand on commodities. It is the uncertainty of the supply capacity of the logistics infrastructure and the logistic activities that arises loss in commodities, relying on which the logistics supply network with capacity uncertain is built and analyzed. Stated above, the optimization model of streamline network of random supply and demand is established, furthermore being proved to be a convex optimization model, which can be converted into a related variational inequality problem. Then modified random projection algorithm is designed. The model is verified and analyzed, drawing a conclusion that the specific logistics service offered according to concrete logistics demand would contribute to the decrease of the loss aroused by logistics service.Finally, the paper studies the optimization model of streamline supernetwork of multi-cycle demand. In consideration of the complex relation between logistics route and logistics chain and that between logistics chain and Origination-Destination, the optimization model is built with analyzing the characteristics of multi-cycle supply and demand network, maximizing matching degree of the generalized cost of supply and demand based on multi-cycle logistics. In case of the streamline supernetwork of the automobile’s making and selling, the matching degree of every cycle is solved and the related logistics solutions is given out according to the logistics demand in every cycle. |