@PhdThesis{duepublico_mods_00041343,
  author = 	{G{\"u}ller, Mustafa},
  title = 	{Optimal Inventory Control and Distribution Network Design of Multi-Echelon Supply Chains},
  year = 	{2016},
  month = 	{Jun},
  day = 	{02},
  abstract = 	{Today, most companies have more complex supply chain networks in a more volatile business environment due to global sourcing, outsourcing of production and serving customers all over the world with a complex distribution network that has several facilities linked by various activities. More companies involved within the value chain, means more nodes and links in the network. Therefore, globalization brings complexities and new challenges as enterprises increasingly benefit from global supply chains. In such a business environment, Supply Chain (SC) members must focus on the efficient management and coordination of material flow in the multi-echelon system to handle with these challenges. In many cases, the supply chain of a company includes various decisions at different planning levels, such as facility location, inventory and transportation. Each of these decisions plays a significant role in the overall performance and the relationship between them cannot be ignored. However, these decisions have been mostly studied individually. In recent years, numerous studies have emphasized the importance of integrating the decisions involved in supply chains. In this context, facility location, inventory and transportation decisions should be jointly considered in an optimization problem of distribution network design to produce more accurate results for the whole system. Furthermore, effective management of material flow across a supply chain is a difficult problem due to the dynamic environment with multiple objectives. In the past, the majority of the solution approaches used to solve multi-echelon supply chain problems were based on conventional methods using analytical techniques. However, they are insufficient to cope with the SC dynamics because of the inability to handle to the complex interactions between the SC members and to represent stochastic behaviors existing in many real world problems. Simulation modeling has recently become a major tool since an analytical model is unable to formulate a system that is subject to both variability and complexity. However, simulations require extensive runtime to evaluate many feasible solutions and to find the optimal one for a defined problem. To deal with this problem, simulation model needs to be integrated in optimization algorithms.</br>
In response to the aforementioned challenges, one of the primary objectives of this thesis is to propose a model and solution method for the optimal distribution network design of an integrated supply chain that takes into account the relationship between decisions at the different levels of planning horizon. The problem is formulated with objective functions to maximize the customer coverage or minimize the maximal distance from the facilities to the demand points and minimize the total cost. In order to find optimal number, capacity and location of facilities, the Nondominated Sorting Genetic Algorithm II (NSGA-II) and Quantum-based Particle Swarm Optimization Algorithm (QPSO) are employed for solving this multiobjective optimization problem. Due to the complexities of multi-echelon system and the underlying uncertainty, optimizing inventories across the supply chain has become other major challenge to reduce the cost and to meet service requirements. In this context, the other aim of this thesis is to present a simulation-based optimization framework, in which the simulation is developed based on the object-oriented programming and the optimization utilizes multi-objective metaheuristic techniques, such as the well-known NSGA-II and MOPSO. In particular, the proposed framework suggests a great utility for the inventory optimization problem in multi-echelon supply chains, as well as for other logistics-related problems.},
  url = 	{https://duepublico2.uni-due.de/receive/duepublico_mods_00041343},
  file = 	{:https://duepublico2.uni-due.de/servlets/MCRFileNodeServlet/duepublico_derivate_00041495/Guller_Diss.pdf:PDF},
  language = 	{en}
}