Моделирование адаптивной самоорганизации экосистем

A method for overcoming the problem of "curse of instability" models of complex systems, which is based (by analogy with natural systems), on the idea of multidrug resistance is offered in the work. On a sample of the model of closed ecosystem proposed by the authors, the possibility of dynamic formation of multiple stationary states (invariants manifolds) in which the system can transit as conditions change. The number of such states depends on the complexity of the system and it is growing faster than exponentially when the system is becoming more complicated. The mapping of this property of complex natural systems is practically impossible with manual design of attractive landscapes, and is based on the methods of the Concept of Adaptive Self-organizing of complex systems (CAS) proposed by the authors. The methods considered in this paper are based on the use of neural network algorithms proposed by the authors, which are used to simulate the adaptive feedback loop. The methods are transitional between the classical methods of modeling and models of self-organizing networks (as an example, it can be networks of interconnected organisms in an ecosystem), proposed earlier by the authors. The introduction of mechanisms of adaptive self-organization of living system, implemented at the level of its elements (for example, for an ecosystem it will be organisms) allows not only to obtain realistic sustainable models, but also significantly improves the quality of modeling and prediction of the behavior of ecosystems and the biosphere on the whole.