Modeling Method Of Distributed Computing Resources For Edge Systems

The variety and volume of computing duties have dramatically increased as a result of the rapid development of mobile communication technology.The requirements for job delay and device energy consumption cannot be met due to the lack of resources in end devices.The industry has suggested the mobile edge computing architecture to address this issue.This architecture offloads device tasks to edge servers for execution,reducing task processing latency and task computing energy consumption.To completely utilize the efficiency of mobile edge computing,however,it is essential to to construct an edge system model and design a reasonable resource allocation and task deployment strategy due to the limitations of the computing resources and storage space of edge servers.This paper offers a deep reinforcement learning-based solution to the issue of reducing energy consumption and delay in resource-constrained mobile edge computing systems.Here is the primary piece of work:(1)The task deployment,edge caching,and resource allocation strategies are jointly optimized to minimize the overall energy consumption of the mobile edge computing system composed of multiple mobile devices and multiple edge servers with various integrated containers,subject to the constraints of container type,transmission power,delay,task offloading,and deployment strategy.The solution to the issue is formulated as a combinatorial optimization problem with numerous discrete variables by modeling the edge system.Simulated experiments show that a deep Q-learning-based efficient edge caching and task deployment strategy(DQCD)is superior to other algorithms at reducing system energy usage and strategy response latency.(2)In the case of multiple mobile devices and multiple edge servers,the edge caching and partial offloading strategies are jointly optimized to reduce the overall delay cost of the system while taking into account the constraints of device mobility,distributed execution of single tasks,limited capacity of edge caching,and limited computing resources of mobile devices and edge servers.The issue is defined by modeling the edge system.A deep deterministic policy gradient(DDPG)algorithm-based edge caching and partial offloading strategy(DDCO)is suggested,and simulation results demonstrate that this algorithm can effectively reduce system delay when taking into account its continuous action space.The system is modeled in various mobile edge computing scenarios in this paper,and the enhanced algorithms based on deep reinforcement learning are suggested to address the issues of minimizing the system’s overall energy usage and delay.Through simulation evaluations,the model and algorithm’s viability and efficacy are confirmed.