An integrated production-logistics approach to partner-chain design in agile manufacturing

The concept of Agile Manufacturing has emerged in response to challenges that companies fare in today's competitive environment, such as the speed with which products are designed, manufactured and distributed, the shortened product life cycles as well as the need for higher efficiency and lower operational costs, In this environment, the success of individual companies is in a major part, tied to their ability to form alliances with partners that can complement their capabilities. A Virtual Enterprise is formed when several companies engage in a collaborative effort to respond to a rising market opportunity. There are three main aspects of Virtual Enterprises: the cooperation between member companies in terms of information sharing, the integration of operations since member companies will act as a single entity, sharing risks and benefits, and the temporal aspect of the alliance, subject to the time frame of the market opportunity.; This dissertation presents the formulation of the problem of formation of Virtual Enterprises given a set of available options, based on the integrated optimization of logistics and production costs along the supply chain they will form. The methodology presented deals with the three main aspects of Virtual Enterprises mentioned above. The integration of companies into a single entity is considered in the integrated analysis of the operation of the virtual organization. The collaboration of partner companies, is implied in the sharing of information about production capabilities, demand patterns and costs. Finally, the methodology presented also takes into account the temporal aspect of the formation of Virtual Enterprises, by providing the coordinator of the Organizational Web with an analytical tool to support the decision making process and allows for analysis of different demand scenarios.; Considering the high degree of collaboration between member companies, this dissertation shows that an integrated analysis better reflects the behavior of the system. Research on analyses of integrated production-distribution systems is still fragmented and the benefits of such analyses has not been explored in depth. The problem presented in this dissertation and its solution show that there can be a significant cost advantage if an integrated approach is taken when solving the problem. The increase in the problem complexity that the integrated analysis presents, limits the solution methodology to heuristic procedures. Two heuristic procedures are presented for the solution of the problem. The first one is based on a k-shortest paths algorithm with an iterative consideration of production capacity dependencies along the supply chain that member companies will form, this heuristic has the advantage of providing alternative solutions as well. The second heuristic is based on a Column Generation methodology, which is a more structured approach to the solution of the problem and proves to be more stable and capable of solving problems where large networks are considered.; A Benders partitioning algorithm is developed to obtain a lower bound for the solutions from the proposed heuristic methodologies. This procedure proves to give acceptable bounds and indicates the accuracy of the heuristic methodologies. Finally a simulation analysis is presented of the impact of demand variability in the solutions obtained.