Joint Scheduling Strategy Of Mobile Computing And Local Multi-core Resources

With the development of 5G,mobile devices equipped with multi-core processors need to handle increasingly complex tasks for applications.Although mobile devices offer considerable convenience through these applications,its have to deal with increased energy consumption and higher delay.However,the computing and energy resources of mobile devices are limited.Particularly,battery capacity is limited by the physical size of mobile devices;this limitation worsens the user experience.Therefore,how to use the limited resources of mobile devices to efficiently perform the tasks is an urgent problem.In light of this,mobile computing(MC)has emerged,which provide remote cloud services for mobile users.The intensive tasks can be sent to the cloud servers for processing due to its powerful computing capacity,which avoids the massive delay and energy consumption problem of the local computing.Additionally,MTC is similar to MC and has the advantage of extending the battery capacity and computing ability of mobile devices,and currently multi-core mobile devices are very popular.Therefore,consider combining MTC and MC to study how to efficiently schedule multi-core resources and edge computing resources of mobile devices.This thesis designs an offloading algorithm for joint scheduling of multi-core and cloud resources for multi-thread dependent tasks.Multithreaded dependent tasks are tasks that can be calculated in parallel and have dependencies between subtasks.Based on the complexity of the dependencies,we have designed two computing offloading strategies,a multi-threaded mobile cloud computing offloading strategy and a fine-grained offloading strategy for multithreaded applications.Both of them jointly schedule the local multi-core calculation frequency,uplink transmission power,and offloading decision.The multi-threaded mobile cloud computing offloading strategy solves the computational offloading problem of complex dependent tasks.Due to the complex dependency relationship,only the suboptimal solution can be found according to the proposed strategy.Precisely,first,formulate multi-threaded computing principles to reduce computing energy consumption or latency.Then,according to these principles,complex dependent tasks are reduced to serial tasks.Finally,a combination of convex optimization method and backtracking algorithm is used to solve the optimal scheduling solution for serialized complex dependent tasks.Simulation results show that compared with the local MTC strategy and the single-threaded offloading strategy,combining MTC with MC significantly reduce the computing energy consumption of multi-threaded applications and improve the computing capacity.The multi-threaded application's fine-grained offloading strategy solves the computation offloading problem of simple dependent tasks.Due to the simple dependency relationship,the optimal solution for the joint scheduling problem can be found through the proposed strategy.Precisely,we first formulate the minimum energy consumption problem for ST offloading.Then,we prove the problem is convex and solve it by standard convex optimization technique.Thirdly,the optimization objectives are extended from ST applications to multithread(MT)applications,and calculation rules of MT applications are designed to reduce computing costs.Finally,based on these calculation rules and the optimal solution for ST offloading,we develop an MT offloading strategy to solve the computation offloading problem of MT applications.Simulation results show that the proposed fine-grained MT offloading strategy effectively reduces the minimum delay requirement of mobile computing.