Joint Resource Allocation In Wireless Cellular Networks With Mobile Edge Computing

Since the recently emerging mobile applications have posed significant demands not only on high data rate but also on high caching and computing capabilities,the growth in communication capability alone is no longer sustainable for wireless networks.The integration of networking,caching,and computing functionalities into one system can provide not only native support for highly scalable and efficient content retrieval,but also powerful capability of data processing,hence reducing duplicate content transmissions and enabling swift executions of computationally intensive tasks.Despite the bright prospect of integrated networking,caching,and computing systems,a number of significant research challenges remain to be addressed prior to widespread deployment of the integrated systems.This dissertation leverages the theories and methods of resource management optimization to study the multi-dimensional joint resource management of the integrated system,aiming at addressing several important issues of the integrated system: joint optimization of communicational and computational resource,computation offloading and optimal resource allocation,joint optimization of caching decision and computation offloading.The distinctive features of this dissertation are as follows.1.The design of a framework for joint optimization of computation offloading and interference management in wireless cellular networks with Mobile Edge Computing(MEC).In order to minimize the overall system overhead in terms of computation task execution time and energy consumption,this dissertation proposes an integrated framework for computation offloading and interference management to offer optimal offloading decision and mitigate interference during offloading process.In this integrated framework,the computation offloading decision,physical resource block(PRB)allocation,and MEC computation resource allocation issues are formulated into an optimization problem that aims at minimizing the overall overhead.First,the MEC server makes the offloading decision in accordance with the load estimation and the principle of overall overhead minimization.Second,an improved graph coloring method with low complexity is employed to obtain the optimal PRB allocation strategy and the optimal frequency reuse parameter.At last,on the basis of the optimal offloading decision and the optimal PRB allocation strategy,the computation resource of the MEC server is distributed to the user equipment(UE)with the goal of minimizing the computation task execution time.The MEC computation resource allocation is achieved via solving two convex problems,which emphasize the minimization of overall cost and the system fairness,respectively.Theoretical analysis and simulation results demonstrate that the proposed solution can efficiently reduce the overall system overhead with relatively low complexity.2.The joint optimization of computation offloading,caching decision,and spectral and computational resource allocation in wireless cellular networks with MEC.Since the conventional resource allocation approaches separately address the communicational,caching and computational resource,this dissertation jointly formulates the computation offloading decision,resource allocation and content caching strategy into one optimization problem,aiming at maximizing the system utility.The modeling of this problem jointly considers the impact of the available spectrum and the amount of computational resource on the offloading decision,the mutual influence of the spectral and computational resource allocation,and the impact of the caching decision on the system revenue,so as to guarantee the maximization of the utility and the holistic balance of the multi-resource allocation.Due to the existence of binary variables and product terms of the variables in the original problem,the problem is non-convex and therefore is hard to solve.This dissertation converts the problem into a convex problem via variable relaxation and substitution of the product terms.The proof of the convexity of the converted problem is given.Furthermore,in order to overcome the drawbacks of high complexity and high signaling overhead brought by centralized algorithms while solving this problem,the original problem is decomposed into a number of subproblems and an alternating direction method of multipliers-based(ADMM-based)algorithm is developed to solve the optimization problem in a distributed manner.Theoretical analysis and simulation results demonstrate that the proposed solution can achieve high system revenue while maintaining fast convergence property and low computational complexity.3.On the basis of the previous research efforts,in order to improve the efficiency of the backhaul and caching resource of the integrated system in video transmission scenario,the dissertation proposes to study the joint optimization of caching strategy,partial computation offloading and resource allocation scheme for video caching and transcoding based on integrated systems.Specifically,the dissertation develops a multi-timescale optimization approach in which the long-term video caching strategy is decoupled from the short-term resource allocation scheme.First,the proposed optimization approach addresses the long-term video caching strategy.Taking into account the uncertainty of the user request arrival in video caching decision and the wasted caching space caused by constant traffic constraint in traditional methods,this dissertation formulates the video caching decision as a robust optimization problem.Aiming at maximizing the caching utility,this robust optimization problem employs the probability-based network traffic constraint and the caching space constraint,so as to guarantee the tolerance of the uncertain optimization parameters and exploit the storage space up to the hilt.Second,the short-term resource allocation for video transmission and transcoding computation is conducted on the basis of the optimal long-term video caching decision.Aiming at the maximization of the system revenue and leveraging the partial computation offloading mechanism,the partial computation offloading strategy and the spectral and computational resource allocation schemes are jointly formulated into an optimization problem.In order to reduce the computational complexity and signaling overhead,a distributed algorithm is utilized to solve this problem.Simulation results demonstrate that the proposed solution can significantly increase the caching space utilization and achieve higher system revenue compared to baseline solutions.At last,the dissertation concludes with an open discussion that highlights the broader perspectives of the integrated system of networking,caching and computing.