Global Optimization of Minority Game by Smart Agents

TL;DR

This paper introduces a new minority game model with smart agents that use trial-and-error to achieve near-optimal global performance, demonstrating self-organization and improved profits over traditional agents.

Contribution

The paper presents a novel minority game model with smart agents employing inductive learning, leading to minimized system loss and enhanced agent profitability.

Findings

Smart agents achieve theoretical minimum variance in the system.

Smart agents outperform noise traders and conventional agents in profit.

The system self-organizes into a highly optimized state.

Abstract

We propose a new model of minority game with so-called smart agents such that the standard deviation and the total loss in this model reach the theoretical minimum values in the limit of long time. The smart agents use trail and error method to make a choice but bring global optimization to the system, which suggests that the economic systems may have the ability to self-organize into a highly optimized state by agents who are forced to make decisions based on inductive thinking for their limited knowledge and capabilities. When other kinds of agents are also present, the experimental results and analyses show that the smart agent can gain profits from producers and are much more competent than the noise traders and conventional agents in original minority game.