A GAME-THEORETIC APPROACH FOR THE PROCESS OF A MULTI-NODE DECENTRALIZED COMPUTING SYSTEM SIMULATING

Abstract

В статье рассматривается подход к имитационному моделированию процесса функционирования многоузловой децентрализованной вычислительной системы, реализуемой совокупностью взаимодействующих вычислителей на базе мобильных терминалов. The article considers an approach the process of functioning a multi-node decentralized computing system simulation implemented by a set of interacting computers based on mobile terminals. As a methodological basis for such simulation, is substantiated a game-theoretic approach implemented in the framework of reinforcement learning methods in multi-agent systems. A formalization of this process is given, taking into account the possibility of increasing the power of the set of possible actions of an individual compute node due to the cooperative and/or competitive interaction of a subset of compute nodes with each other. An approach to obtaining criteria for dynamic stability of a multi-node decentralized computing system used training of compute node the Nash equilibrium point is proposed.