Article Preview
TopIntroduction
In a recent works (Riahla et al, 2014; Ait-Salem et al,2012), we proposed a hybrid method entailing proactive and reactive methods and based on the ant behavior (Dorigo et al, 2006; Dorigo et al, 1999; Di Caro et al, 1998; Singh et al,2020; Tseng et al, 2021; Okwu et al, 2021; A.C.J., Malar et al, 2020). The classic ant routing protocol (Correia et al, 2009; Correia et al, 2008; Bouazizi, 2006; Laxmi et al, 2006; Caro et al, 2005; Di Caro, 2004) use a broadcasting method that exponentially increases the routing overhead. In order to bypass this weakness, we introduced a new idea through an ant-based algorithm. The main idea consists of using Ant-Agents in order to find optimal routes in the network. These Agents are generated by each network node. When a new connection is planned to be established between nodes n and d without a corresponding entry in n’s routing table, this node n creates a new route request. Unlike other protocols, the node n set a local route request (rather than broadcasting it) available for Ant-Agents passing through it. It is, therefore, the role of these Ant-Agents that move within the network to 'intelligently' share this information in the network and to provide routes towards a destination node d. These Agents will disappear from the network whenever they accomplish their tasks or their lifetime is over. Each Ant-Agent will be created only if its predecessor is eliminated. Consequently, the number of Ant-Agents available in the network will be the same as the number of nodes. As a result, this approach will decrease the routing overhead.
The whole proposition performances were then evaluated using simulations. The obtained results showed a significantly improvement in terms of overhead, end-to-end delay and packets loss ratio.