Modeling of the network design of a third party logistics company under uncertainty: Stochastic modeling and genetic algorithm implementation

The difficulties of competition in recent years and the rapid development of technology have increased the importance of issues such as improving customer service level, reducing production times and delivery times, and reducing stock levels. Firms have started to work on issues such as supply chain management, logistics management, production distribution planning to reach the necessary goals in these issues. The "Reverse Logistics" issue, which takes into account the return and recovery of products, has also become an area of much work over the last few years. This study includes the logistics network design concept which has been increasing in recent years. At the same time, it is a study that takes into account not only forward but also backward flows. In this study; five shipping companies, two cargo collection centers, two airports, two delivery / distribution centers and four buyer shipping countries are investigated. At the same time, the study includes reviews of uncertain cargo volume, periodic cargo status, and collection and supply branch situations and delays of opportunity loss and reliability percentage of centers so that and what percentage of deliveries are sent smoothly, at the right time and in the right way. The purpose of this study is; the assessment of whether current collection and delivery centers are open or closed, and the determination of what amounts to be sent between centers. Thus, the study will provide information on whether the collection and delivery centers are open or closed and the flow between centers, airports, senders and receivers to the operation subject. It is also aimed to set an example for scientists working on the logistics topic, and for those interested in a network structure. The data in the study are obtained directly from the cargo company and the problem is first modeled by using stochastic programming. The stochastic network design model created as a maximization problem is solved using the GAMS Distribution 23.5.1 program. Then, since it is known that every scenario in the stochastic model has equal probabilities, the average of the demand quantities included in the scenario is taken and for the deterministic model it is assumed that the assumption of this data is realized and two models aiming at a profit maximization and cost minimization solved using GAMS Distribution 23.5.1 program. As a third method, a genetic algorithm based model which is a metasequential method has been created. After the results were obtained, the three methods were compared to each other in duplicate.