Research On Multi-agent Cooperative Control Method In Mobile Wireless Sensor Networks

Mobile Agent technology is a wireless sensor network technology to achieve a new kindof intelligent. According to the trajectory, mobile agent which sent by the data processingcenter can visit source node of a network automatically to collect the data for compressionand integration. Compared with the traditional C/S mode, it can reduce the redundant dataand the amount of data flowing in the network, save the network bandwidth, and providehigher energy efficiency and lower delay of network. With the research of the multi-mobileagent technology developing, the performance of network is getting better and better,especially for those large-scale networks. However, with the continuous expansion of thenetwork, each mobile agent needs to visit a lot of source nodes. As the data packet size theagent carried is growing, it will take too much energy for the data transmission. Based on themobile agent in wireless sensor networks, the trajectory of the mobile agent optimalcontrolling is one of the key issues of network performance optimization. Its performancedirectly determines the transmission speed of information and energy depletion. The mainwork is as follows：(1) In this paper, combining with the idea of the minimum spanning tree algorithm andGCF algorithms, we propose a new multi-agent trajectory collaborative optimization methodfor nodes to marshalling planning reasonably and efficiently, making the trajectory of mobileAgent leaving from the data center, moving to the visited nodes, and then back to the datacenter joint best and improving the efficiency of the implementation of the entire task flowmulti-agent. Compared with the traditional minimum spanning tree method, this method hasthe advantage that it not only solved the long latency problems of the ordinary single mobileAgent tasks, while avoiding the excessive tree branches which led to the formation of nodesin large quantity and wide distribution problem. Finally, the simulation of the proposedmethod and the minimum spanning tree method were compared, the results show that thismethod is more superior in the task delay and energy consumption.(2) In this paper, considering a tree structure which based on route planning algorithmfor mobile agent, we propose a multi-mobile agent cooperative control technology for dataseparation. According to data distribution control method, the data packages between themultiple mobile agent and migration process in each mobile agent are reasonably allocated,which solve the excessive consumption when agent carries large a mounts of packetstransmitted on the network. In addition, it would reduce the delay time and shorten theworking life of the source node. Finally, the simulation of the proposed method and theminimum spanning tree method showed that the proposed method has better performance inenergy consumption task latency and the source node working life.