| Regional Communication Network (RCN) is composed of self-organized mobile nodes and wireless channels. It is an important part of Tactical Communication Network in military communication system. In order to ensure the fast and reliable transmission, there are several requirements for the routing algorithm, such as adaptive to the dynamic changed network topology, fast route setting up, dealing with unidirectional links, loop-free, and so on. Furthermore, the energy balance is especially important because the energy of each node is finite.Ant colony optimization algorithm is a kind of swarm intelligence inspired by the behavior of real ants, which has been widely applied to many combinatorial optimization problems. Ant colony optimization algorithm is particularly suitable for RCN routing problem because it supports distributed computation and the single agent is easy to be implemented.In this thesis, a novel routing algorithm based on ant colony optimization for RCN is proposed. It is a hybird ant routing algorithm composed of proactive routing and reactive routing. The convergence of the algorithm is proved, and the scopes of parameters are analysed which is useful in running the simulation. The proposed ant routing for RCN has got some achievements described below.In the first place, in order to extend the survival time of the entire network, the thesis uses energy as a heuristic factor, so that the forward ant prefers to choose path with higher residual energy. Meanwhile, the backward ant updates the incremental pheromone according to the value of the paths' total residual energy. The simulation results show that the energy balance performance is more excellent than AODV.Secondly, the topology of RCN is dynamic as its nodes have high mobility. A new established route may be invalid at any time, which would result in high packet loss ratio and high retransmission ratio. The thesis considers the distance of two nodes as a metric of the links' stability. During the establishment of the route, the forward ant selects the more stable links in high probability. The probability is calculated by the formula in which the link stability is involved as a heuristic factor. The new route would have a long survival time by this way. And the thesis proposes a mechanism of link failure detection and route repair in this situation. The simulation results show that the average delay is lower than AODV.Thirdly, every node's transmission range is generally various due to the affection of terrain and transmit power difference, which makes a lot of unidirectional links. In order to solve this problem, each node detects the existence of unidirectional link while updating neighbor table periodically. The nodes which contain a unidirectional link will be removed from the neighbor table. Then, the backward ants can return to the source node in the path of the forward ants. The efficiency of the routing algorithm is greatly improved with this method.Finally, a new ant routing protocol named AntRoute is implemented in the network simulator NS-2. Then, this thesis compares the performance of AntRoute with AODV through a lot of simulations in different environments. Results show that the AntRoute's energy balance is better than AODV. It can be concluded that the ant routing algorithm is more appropriate for RCN. |
PDF Full Text Request