The Algorithm Of Computing Migration Of Inspection Robot Based On MEC

The digitization and intelligence in modern industrial scenes are developing very rapidly.As all kinds of artificial intelligence-based service applications are implemented in industrial scenes,it poses a challenge to the computing resources of industrial systems.In order to meet with the increasing demand for computing resources year by year and alleviate the load pressure of the cloud,mobile edge computing technology is widely used in industrial computing scenarios to provide high-efficiency,high-performance computing migration services for industrial equipment.In this paper,based on the background of industrial intelligent inspection,combined with the characteristics of industrial intelligent inspection,the computing migration algorithm of the inspection robot is designed.Considering the mobility of the inspection robot,find out the optimal computing migration strategy for it,so as to minimize the energy consumption of completing the inspection task during the inspection,and extend the standby time of the inspection robot.Reduce inspection costs.Specific work include:（1）In the single inspection robot inspection scenario,the industrial intelligent inspection scenario is analyzed,the robot inspection movement model and cycle migration model are abstracted.The task is modeled as directed acyclic graph,and the task energy consumption minimization optimization problem is established.A constraint decomposition algorithm is proposed to decompose the delay limit and migration path limit of the total task into the constraints of each sub-task according to the characteristics of the task.The total task energy minimization problem is decomposed into each sub-task energy minimization problem.Finally,the optimal MEC node is selected for each sub-task with greedy thought.Simulation results show that the algorithm is effective.（2）In the scenario of multi-inspection robot,this paper designs a MEC node migration model based on the traditional computing migration model.The competition of bandwidth and computing resources among multiple inspection machines are considered.To minimize the average energy consumption of all inspection robots,find the best MEC node for each inspection robot to upload tasks and execute tasks.A distributed algorithm based on game theory is designed for the two stages of upload and execution respectively,and the Nash equilibrium solution of the two stages is found through iteration.The convergence and effectiveness of the proposed algorithm are proved by simulation.（3）Based on the research of the first two parts,this paper also implements a system platform for algorithm simulation verification.The platform provides functions such as edge network,inspection robots and tasks construction,permanent storage,modification and deletion,which is convenient for researchers to reuse existing examples.The platform stores a large number of algorithms that need to be simulated for researchers to choose,and provides the function of result analysis,which greatly reduces the workload of simulation.Researchers can easily interact with the platform through the graphical interface,and can focus more on algorithm design.