[2M4-OS-3a] OS-3 (1)

[2M4-OS-3a-04] Investigation of genetic algorithm for autonomous AI using blockchain

〇Kanta Nanri¹, Yutaka Moriyama¹, Koichi Nakayama¹ (1. Saga University)

Keywords:Blockchain, Autonomous agent, Genetic algorithm

This paper proposed ``Self-Administered Genetic Algorithm (SA-GA)'' as an evolutionary method that applies systems of intelligent autonomous agents. The intelligent autonomous agents were implemented as an autonomous program called a smart contract and executed by themselves according to pay the cryptocurrency to a node of the block chain. The intelligent autonomous agents' copies (children) were created by acquiring the cryptocurrency from their parents. In the examination of this paper, SA-GA was applied to a problem that resources which a large number of unspecified users provided were allocated to a large number of unspecified users. The result showed that SA-GA could obtain an adequate solution without a management for whole system. Moreover, it was shown that these solutions could be obtained in practical calculation time. Namely, SA-GA can be a method that optimizes the system consisting of a large number of autonomous agents.