The thesis presents new approaches to model and solve the dynamic replica placement problem with service levels in content delivery networks. First, a new modeling approach is proposed. The objective is to minimize the total cost of replica storage, placement and delivery subject to service level constraints. The problem is formulated as a mixed integer program which identifies the optimal amount and location of replicas of a single content over several periods and the according transfer paths. Second, new solution methods for the problem are proposed. A simulated annealing approach and several variable neighborhood search variants are presented. Third, the computational study shows the efficiency of the proposed algorithms. The results indicate that high quality solutions can be obtained quickly. Furthermore, in comparison to established heuristics from the literature considerable cost savings are possible with our metaheuristics.
«
The thesis presents new approaches to model and solve the dynamic replica placement problem with service levels in content delivery networks. First, a new modeling approach is proposed. The objective is to minimize the total cost of replica storage, placement and delivery subject to service level constraints. The problem is formulated as a mixed integer program which identifies the optimal amount and location of replicas of a single content over several periods and the according transfer paths....
»