Supply Chain Super Network Design——Definition,Model And Algorithm Improvement

The supply chain network system is the soul and backbone throughout the overall supply chain management. Nowadays, on account of economic globalization, the international co-operation based on division of labor impels the complexity of supply chain network. Supply chain manager has to deal with many decisions. The optimizations of such decisions not only able to enhance competitiveness of the entire supply chain, which brings to boost of its market share, but also save and reduce the supply chain network construction, operation costs, and the whole society high level logistics costs. The physical layer of supply chain network involves facility location, distribution planning and so on, be able to guide the regional logistics planning. Since the realistic supply chain network belongs to supernetwork category. Super-network is an effective tool to describe the complex networks with the multi-layer, multi-dimensional, multi-level, and multi-properties flow..It is more realistic to optimize the complex supply chain network system by means of abstraction, supernetwork-representation, and development of corresponding conversion technique from complex network system into regular network. Such mechanics are able to provide practical planning and operation decision support. Furthermore, such optimization method is the supplement of logistics subject, systems optimization theory, graph and network flows, and complex network theory, is also the new topic of logistics and supply chain research area. Their research results may reveal operating law of both the supply chain and logistics system. Besides, the results is the crucial element of regional logistics planning, supply chain network strategic design, operational planning and decision, as well as the basic research parts of supply chain system optimization theory. The related study is of urgency and necessity.Based on current research, this dissertation provides the systemic concepts and problem definition for supply chain supernetwork design for the first time. Moreover, it contributes to the operation pull-power of supply chain network-logistics demand, with the employment of modern data processing tool, represents the logistics demand in supply chain supernetwork to meet the actual world. Formulates integrated supply chain supernetwork frame, models and algorithms to realistic complex supply chain network. The main contents of this dissertation consist of:1. Firstly, we specify the conception and define the supply chain supernetwork design problem. Since the long-term planning, as well as the operational optimization and design problem of supply chain system are belongs to supernetwork category, we transform complex network into regular network with supernodes and links using the equivalency optimization approach. Further, the corresponding network model and mathematical models are formulated to solve the problem, as well as the design procedure of integrated supply chain supernetwork.2. Secondly, we extent the representation method and theory of logistics demand in supply chain network. Based on classification of logistics space distribution demand into three groups, which are random, period uncertainty and certainty, we propose the concepts of logistics time demand representation. Applying the Arcgis9.3software, we are able to formulate and get the solution of multi-objective facility location problem with logistics random area demand.3. Thirdly, we define the strategic-phase, certainty logistics time demand-based simple supply-demand supernetwork design problem. The decision problem under consideration consists of supply facility location, supply facility capacity design, path-construction, and extension of transportation network. The model formulation merges both minimal system cost objective and maximal customer demand coverage of facility location. In addition, we describe the mechanism of complex network structure transform into ordinary regular network. Develop specific Lagrangian relaxation algorithm, and using computational experiments for sensitive analysis of the costs in the model, identify the trade-off between each part of the cost and total regional construction investment cost.4. Fourthly, similar with the third part, we define the strategic-phase, certainty logistics time demand-based multi-product, multi-layer supply chain supernetwork design problem, and present the minimal costs formulation which can also drives facility location solution to obtain maximum demand coverage. The decision involves consists of manufacturing plant or core enterprise warehouse location, capacity design, corresponding transportation network construction and extension decision, preliminary distribution plan, critical raw material supplier selections and critical customer market selection. With the propose of transform mechanism form supply chain supernetwork topologic structure to regular network topologic, as well as relax the network fixed cost, we develop the improved node potential network simplex algorithm. The experiment with random generation data proves the efficiency of algorithms and models.5. Then, a definition of operation-phase, uncertainty logistics time demand-based multi- product, multi-layer supply chain supernetwork design problem is described. Based on uncover degree, we analyze the supply chain new equilibrium conditions, two supernetwork design model are presented for this case with one account for equilibrium conditions and one without. Using the expected value of objective function equivalent solution approach, a specific Lagrangian relaxation algorithm and a variational inequation iteration algorithm are proposed for the models respectively.6. Finally, a south-central China real case study with the supemetwork design model corresponding to simple supply-demand network and an overall supply chain supemetwork design case of southwest United States for large-scale retailer with the using of entire two-phase model are proposed and solved, The cases for real world point out the theatrical research result possess of practical meaning, is easy to combine with real case problem.