| With the development and progress of intelligent inspection,more and more AI-based services are applied in inspection,and it is difficult for the inspection robots with limited computing and energy storage capacity to meet the current demand.In order to cope with the dramatic increase in computing power demand and relieve the computing load of cloud computing center,this paper applies Mobile Edge Computing(MEC)and Device to Device(D2D)communication technology to the intelligent inspection system.Compared with the cloud computing environment,the resources in the edge computing environment are more limited,so the task offloading strategy plays an important role in the system performance.In this paper,the computational offloading strategy in single-base station multi-terminal and multi-base station multi-terminal inspection scenarios and the virtual machine migration resource allocation strategy in multi-base station multi-terminal inspection scenarios are investigated for the shortcomings in single-base station multi-terminal and multi-base station multi-terminal inspection scenarios,respectively.The existing task offloading methods are more dependent on the central controller,and game theory,with its distributed decision mechanism,can not only effectively reduce the dependence on the central decision,but also balance the interests of each decision subject.The current research on task offloading based on game theory only considers MEC technology,but in this paper,we study the task offloading algorithm based on game theory and applicable to D2 Dassisted MEC system.Secondly,most of the existing studies focus on the resource allocation strategy during the migration of a single VM,but in this paper,we use game theory to study the resource allocation strategy of multi-VM migration,which fully considers the interests of each migration task.The main work of this paper is as follows:1)System model construction of single base station multi-terminal inspection scenario.This paper combines the network parameters,equipment performance,computation and communication parameters under the real intelligent inspection scenario,integrates the traditional cellular communication and multiplex communication,and quantifies the communication and computation model under the intelligent inspection robot scenario in the form of a mathematical formula for the joint task offloading decision and resource allocation problem under the single base station multi-terminal inspection system,with the maximum task completion time as The D2D-assisted MEC system model is established with the constraint of maximizing the weighted sum of task delay and terminal energy consumption as the optimization objective.2)Research on the computational offloading strategy under the single base station multiterminal inspection scenario.In order to make the D2D-assisted computation in the cellular coverage area with the lowest cost for the inspection robots,a novel general graph of D2 D link allocation band power is constructed based on the quantified computation model and communication model for terminal resource heterogeneity and task variability.Based on this,a solution based on graph matching is proposed in this paper.Second,based on the phenomenon of co-channel interference under spectrum multiplexing,this paper proposes an interference threshold construction strategy to ensure the quality of cellular communication,derives the existence of Nash equilibrium for the terminal channel game problem,and designs a distributed channel game algorithm,referred to as DCSA algorithm.Experiments show that the method proposed in this paper has 19.7% and 2.4% cost savings compared with the ant colony and blooming algorithm(D2D distribution algorithm),respectively,in terms of task execution delay and energy consumption weighted sum and other indicators of the inspection robot.3)Study on the resource allocation strategy of virtual machine migration in the multi-base station and multi-terminal inspection scenario.In this paper,a system model under this scenario is constructed,a bandwidth trading theory is proposed,and a beneficial bandwidth trading theory is presented.Further,this paper derives the existence of Nash equilibrium of the game and designs a bandwidth resource gaming algorithm based on contractual game,referred to as GAME algorithm.Experiments show that the method proposed in this paper has 0.64%,2.43% and 33.75%performance improvement over genetic and particle swarm optimization algorithms and simple resource allocation algorithms,respectively,in terms of quality of service and migration time weighted sum and other metrics. |
