Aerospace Instrument-Making Annotation << Back POLYMODEL OPTIMIZATION OF NETWORK CONFIGURATION AND INFORMATIONAL OPERATIONS SCHEDULE, PROBLEM STATEMENT AND SOLVING APPROACHES D.A. Murashov, V.A. Ushakov Modern computational networks are characterized by reliance on virtualization technology. Virtualized environments share technical resources (memory, I/O) of a single physical computational unit, which entails the problem of optimal resource distribution. Among the multitude of classes of networks, there is one that is of a special interest to us called industrial internet networks. There is a particularly interesting feature to this class of networks which is predictability of information processes taking place in them. The high extent of determinism inherent to this class of networks implies the possibility to plan both information processes and confi guration of the network, since the former shape the latter. There is an approach enabling optimization of both the confi guration and the plan. The approach comprises using a two-stage iterative procedure first stage of which performs optimization of the plan considering the confi guration unchangeable, while the latter works in reverse – optimizes configuration under the constraints of a plan formed at the first stage. Two stages run in sequence, until some convergence condition is satisfied. The following paper offers a development of this approach in application to heterogeneous industrial internet networks with changing topology. We propose to frame the first stage as the problem of linear programming, and the second – as the problem of stochastic optimization of an agent-based simulation’s parameters, and provide a thorough formal description of the aforementioned simulation. Keywords: heterogeneous networks, optimization, operation planning, virtualization, simulation. DOI: 10.25791/aviakosmos.8.2022.1293 Pp. 23-32.