Using optimization techniques and meta-heuristics is not very popular for MANETs; in particular, the use of multiobjective optimization techniques such as multi-objective evolutionary techniques is not suitable. Is there any justification for this? And what about the few works that use such techniques already available in the literature?
In MANETs, the topology of the network is dynamic i.e. there are frequent changes in the topology due to node and link failures. The optimization techniques require time to execute. During this time, a further change in the topology may require to recompute and rerun the algorithm or protocol. Also, the information about the whole topology of the network is not available at a node. In other words, a node knows only about its neighbors. Therefore, optimization techniques, though good, are not preferred or advisable for MANETs.
In such applicattions, the use of step-wise algorithms is more convenient.
Like Ash Abbas posted, meta heuristics are computationally expensive, they need a good processor and their processing consumes energy. Many MANETs nodes have slower and cheaper processors, and by definition most nodes have limited energy (battery) available. Local vs global data availability is also a trouble, but only for MANET with thousands of nodes, I think.
many thanks for your replies ! what about the following ideas:
1- using Dynamic optimisation : to deal with mobility issues
2- working on VANETs where nodes have more energy and computing ressources
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Artificial Intelligence