Ana dağıtım üssü kapsama problemi için yeni eniyileme modelleri

Hub location problems are derived from classical origin-to-destination problems. The difference of hub location problems is the existence of hub nodes that merge/split transportation demands. A hub also acts as a node where demand originates or arrives. The presence of hub nodes simplifies the network structure. Because of the economy of scale, the hub networks' benefit is saving transportation costs between hubs. This kind of problem is commonly encountered in the design of postal distribution, airline passenger transport, and telecommunication networks. In the current literature, hub location problems are classified into hub median, hub center, and hub covering. The least studied of these is the hub covering problems due to the network's lack of full connectivity. For this type of problem, the elements of capacity and transportation costs are generally ignored in studies. For filling this gap in the literature, new deterministic multiple assignment hub covering models have been developed. These proposed models were examined with numerical experiments using benchmark data sets. It was revealed how important it is to consider the elements mentioned above in the hub covering problem. Another issue that has not been studied for hub location problems in general and hub coverage problems, in particular, is the effect of uncertainties that may be encountered. In this thesis, new two-stage stochastic mixed-integer linear optimization models have been developed to examine the effect of uncertainties on transportation costs and transportation demand. The L-shaped algorithm has been adapted to solve these models, and numerical experiments have been performed with the help of benchmark data sets. As a result of these experiments' analysis, it was shown that the uncertainties should be taken into account for the hub covering problem.