Research On Resource Allocation Strategy In Virtualized Wireless Sensor Network

As a key component of the Internet of Things,Wireless Sensor Network(WSN)has been widely used in fields such as environmental monitoring,industrial production,smart medical care,and smart cities.However,traditional WSN is oriented to specific fields and established tasks,and its network deployment and services are tightly coupled.For new user services,WSN needs to be redeployed,resulting in low network resource utilization.In order to solve the “rigidity” problem of traditional WSN network,related researchers introduced virtualization into WSN.By abstracting and isolating WSN resources,different users can share WSN's perception resources,providing higher flexibility and scalability,so as to realize sensing as a service.Therefore,how to effectively utilize the limited WSN resources to meet the increasingly diversified business needs is of vital importance.This thesis takes the resource allocation strategy in the virtualized WSN scenario as the research object,and the main research work is follows:Aiming at the problem of effective utilization of virtualized WSN resources,a resource allocation strategy of virtualized WSN based on reinforcement learning is proposed.A utility maximization model for virtual sensor network service providers is established based on the requirements of service sensing information quality and the constraints of resource capacity.Considering the randomness of business requests and the dynamic changes of network environment,the resource optimization problem is transformed into a Markov decision process,and the Actor-Critic reinforcement learning method is used to continuously interact with the virtualized WSN environment to obtain the optimal resource allocation strategy.The simulation results show that the proposed method can effectively improve the revenue of service providers and reduce the cost of virtual sensor network request deployment.Aiming at the problem that the non-uniform distribution of energy consumption of sensor nodes seriously affects the life cycle of the underlying WSN and the request acceptance rate of virtual sensor networks,a virtual sensor network mapping algorithm with balanced node energy consumption is proposed.Comprehensively consider node energy status factors,resource capabilities and topological attributes,establish a node multi-index evaluation model for mapping target node selection.Considering the interference constraints between the underlying wireless links,the mapped path was selected by estimating the path mapping cost according to the current energy state and communication energy consumption of the path.Then,a reconfiguration strategy was designed to minimize the migration cost,and the virtual sensor nodes and corresponding virtual links were dynamically adjusted.The simulation results show that the proposed method effectively improves the virtual sensor network request acceptance rate and network resource utilization rate,balances the energy consumption of the network,and extends the network life cycle.