UBC Theses and Dissertations
Joint resource management and pricing in edge computing Nisha, Tarannum
Edge computing (EC) has emerged as a vital technology that works in tandem with the cloud to reduce network traffic and enhance user experience by distributing computational and storage resources closer to end-users and data sources. Despite the tremendous advancements made in EC technology and the enormous potential it holds, it is still in its infancy stage with numerous open challenges to overcome. In this thesis, we particularly aim to design efficient algorithms for pricing, service placement, resource management, and workload allocation in EC. While considering the joint resource management and pricing problem in EC, we take into account the preferences of the services. Specifically, we propose a novel bi-level optimization framework to assist the EC platform to determine the optimal edge resource prices not only to maximize its profit, but also help each service find an optimal resource procurement and workload allocation solution to minimize its cost while improving the user experience. When there is a single edge node (EN), we derive a simple analytic solution for the underlying problem. However, for general case with multiple ENs, the follower problem becomes sophisticated. To this end, we develop two efficient approaches based on the Karush-Kuhn-Tucker (KKT) conditions and linear programming duality, respectively, combined with a series of linearization techniques to optimally solve the underlying bi-level optimization problem. The proposed optimal solution will maximize the profit of the EC platform and improve the edge resource utilization while minimizing the cost of every service. Numerical results demonstrate the superior performance of the proposed dynamic pricing scheme. When the services need to pay for the service placement costs, the follower problems contain integer variables, which results in an extremely challenging bi-level mixed integer optimization problem. Due to the non-convex lower-level problems, we cannot use the KKT or duality-based approach to transform each lower problem equivalently into a set of linear constraints. Inspired by the column-and-constraint generation method from the adaptive robust optimization literature, we design an iterative algorithm to find an exact optimal solution to the formulated bi-level integer optimization problem.
Item Citations and Data
Attribution-NonCommercial-NoDerivatives 4.0 International