Research Of Task Scheduling Optimization In Wireless Sensor Networks

In Wireless Sensor Network(WSN),the sensors are powered by a one-time battery,and due to the low computing power and the limited energy of the sensor,the sensor can not compelet complex applications alone.Therefor,the gateway needs to decompose the application into multiple priority-related tasks,and schedule these tasks to multiple sensors to process.In Wireless Sensor Network,reducing the energy consumption of task completion,balancing the energy consumption of nodes helps to prolong the life of the network,and the real-time and reliability of task completion are important indicators to evaluate the application's performance of network.Then,The research of task scheduling optimization in wireless sensor networks needs to consider several performance indicators such as time,energy consumption,reliability and energy consumption balance.In addition,the failure of nodes in the network leads to the dynamic changes of the environment.It is also crucial to study the dynamic detection and dynamic scheduling of the changed environment.In this thesis,the technical characteristics of task scheduling model and algorithm in WSN are analyzed and elaborated in detail.In terms of wireless sensor networks deployed randomly in this thesis,a multi-objective model of task scheduling for wireless sensor networks is established,and a multi-objective task scheduling method based on energy balance which called MTSEB is proposed.The algorithm uses the Pareto-based multi-objective optimization method to get the optimal solution set of task scheduling,and the indicator of energy consumption balance is firstly taken as the decision index of the final solution,then,the individual with the most balanced energy consumption in the Pareto optimal solution is selected as the final solution.In this thesis,by comparing with the existing algorithms,the proposed algorithm has the shortest task completion time and the highest reliability of task completion under the condition of the typical number of tasks being 20.Simultaneously,the optimized task completion energy consumption ratio ITAS is reduced about 28%,which proves the superiority of this algorithm in reducing the energy consumption of the task.Considerring the dynamic change of network topology in WSN,a dynamic task scheduling model is established and a XOR-based dynamic environment generator is firstly used to simulate the dynamic change of task scheduling environment.This thesis proposed a dynamic task scheduling algorithm based on memory which called DTSAM,it can track the most optimal task scheduling strategy in real time.The algorithm detects environment changes in each generation and updates the node state in real time.If it detects that the assigned node fails,the algorithm jumps out of the current optimal state and adjusts the search space in the next generation according to the constrained,optimizes the task scheduling strategy dynamiclly.Simulation results show that when the task scheduling environment changes every 100 generations,the proposed algorithm can reduce the ? value by about 17%compared with the EIGA algorithm,which proves the proposed algorithm can detect the dynamic changes of environment in time and can effectively track the optimal strategy of task scheduling.