AI Bibliography

Self-organizing flying ad-hoc networks are increasingly used to solve the tasks of recording, storing and transmitting data in space. In some cases, such networks are implemented on a hierarchical basis (mesh topology) and are specialized. These networks are rather volatile since the failure of one of the nodes can disrupt the entire network, and the time it takes to reconfigure the network may be too long. The survivability of the network is an important aspect of the main function (goal achievement) performance reliability. The network has to perform the main function throughout the operating time. During and after the impact of adverse factors, in order to perform the main function, the network has to restore its functions in a minimum time. A characteristic property of a natural or techno-productive negative influence is its low predic- tability, suddenness, instantaneous distribution, the chance of damaging net- work nodes, low probability of failure of the network nodes outside the scope of the factor. The goal of survivability is to perform a target function in the event of a malfunction or network failure and the possibility of complete timely recovery in the case of a failure. The article is devoted to the development of a method for ensuring the survivability of flying ad-hoc network. Effective ways to ensure the survivability of the network in adverse conditions is the application of reconfiguration scenarios, redistribution of functions in the network among nodes, temporary self-isolation of nodes, etc. The proposed method is based on the use of the strategy of structural and functional network reconfigu- ration. This strategy is based on the aggregate-decomposition approach to net- work nodes. Experimental studies show that the probability of maintaining the functionality of the network when using the strategy of structural and functional reconfiguration increases the probability of performing the main function during the influence of the negative factor up to 15% and after it - up to 45%. The analysis of the obtained results shows that additional experimental research is needed to accumulate statistical information for modification of the method in the context of the introduction of self-learning elements.