Optimization Of OLSR Routing Protocol In Large-scale Mobile Ad Hoc Networks

Mobile Ad hoc Network(MANET)is a self-organizing,self-managed,non-center multi-hop wireless autonomous network that does not require other infrastructure support,and has the advantages of rapid deployment and easy expansion.At present,MANET are mainly used in wireless sensor networks,military fields,and emergency occasions.MANET is mainly used in the battlefield environment in the future,so not only the number of network members has increased,but also the nodes have high mobility,which has an impact on the stability of the network,resulting in a decline in network performance.However,existing routing protocols in mobile ad hoc networks have some problems such as increased routing overhead in large-scale networks.And due to dynamic changes in network topology,end-to-end delays increase,the network performance decreases,so they cannot meet the needs of large-scale MANET.Therefore,it is very important to design a suitable routing protocol for large-scale networks.Based on the Optimized Link State Routing Protocol(OLSR),this article optimizes OLSR performance,combines clustering ideas,cross-layer optimization methods,and reinforcement learning algorithm Q-Learning to design a routing protocol CQLOLSR(Q-Learning based Clustered OLSR)to fit large-scale scenarios.Compared with the standard OLSR routing protocol,firstly,CQL-OLSR routing protocol uses a clustering structure that combines the design of clustered routing and flat routing.The control messages of the routing protocol are no longer flooding the entire network topology.Instead,OLSR is used to forward broadcast packets in inter-cluster,and only a few gateway nodes forward cluster link messages between two clusters.So it can reduce routing overhead.Secondly,combine the cross-layer optimization method to reduce the overhead of HELLO messages in the OLSR routing protocol in the cluster.Thirdly,in order to enhance inter-cluster communication capabilities,a dynamic gateway election method is designed.It selects suitable gateway nodes for dynamic network topology,enhances the stability of large-scale network topology,and guarantees the communication capacity of the network.Then propose a stability MPR selection algorithm to further reduce network routing overhead and improve network performance.Finally,based on the idea of Q-Learning,a new routing path selection strategy is proposed in consideration of node mobility and link rate to improve network stability and service quality.The feasibility of CQL-OLSR routing protocol is verified by OPNET simulation software.And by comparing with the standard version of OLSR performance,verify the impact of routing protocols on network performance under different scales,gateway node movement,and different routing strategies.With reference to the results of routing overhead,end-to-end delay,and service reception volume,it is verified that CQL-OLSR can effectively reduce routing overhead and improve network scalability under the premise of ensuring network performance in a large-scale mobile ad hoc network.