Research On Task Allocation And Switch For Dynamic Coalition In WSN

In Wireless Sensor Network (WSN), a single node can’t accomplish system task independentlyfor the strict limit of node energy, processing power and communication bandwidth. Thereby thecooperation work in multiple nodes is needed, and the task allocation becomes the key problem.Based on limited resources, task allocation and nodes schedule are optimized by WSN task allocationtechnology, which has the decisive impact on system performance and quality of service.Consequently, this makes the research of task allocation turn into the current hot spots in WSN.In this paper, WSN task allocation mechanism is studied from two aspects:On the one hand, this paper focuses on the collaborative task allocation mechanism of targettracking in wireless sensor networks. Aimed at the optimization defects of the general task allocationalgorithms which deal with one single objective, we propose a new task allocation algorithm based onmulti-objective optimization. Firstly, an optimization model based on a dynamic coalition isestablished with multiple objectives like tracking accuracy, energy consumption and load balancing;Secondly, the multi-objective evolutionary algorithm NSGA-II is adopted to deal with the aboveoptimization model; Finally, a novel decision-making strategy based on the degree of compromise ispresented to decide the final task allocation scheme from the set of Pareto-optimal solutions.Simulation results indicate that the algorithm optimizes multi-objectives in parallel, converges to theglobal optimal solution with high speed, and achieves better overall system performance than thegeneral task allocation algorithms; the validity of the model and algorithm is verified.On the other hand, the traditional discrete formation of dynamic coalition costs large energyconsumption and is easy to lose historical information. A new smooth handover method for dynamiccoalition is put forward to improve the above defects. The smooth transition of coalition is realized byprocessing a series of actions, which contain: the estimate of target location, the check of switchingtrigger, the selection of coalition head, the switch between former and new coalition head and thealteration of coalition members. Moreover, the repair mechanism is proposed in case of the track loss.This method will help to avoid the repeating formation and keep updating historical informationwhich the former coalition collected. Compared with the traditional method, this new method costslower energy and less network communication and improves the precision of tracking.