Adaptation using hybridized genetic crossover strategies

TL;DR

This paper introduces a hybridized genetic crossover method for agents in a game to adapt strategies dynamically, inspired by biological evolution, aiming to improve their performance and survival.

Contribution

It proposes a novel hybridized genetic crossover mechanism for strategy adaptation in agents, inspired by biological evolution, and evaluates its effectiveness in a simulated game environment.

Findings

Agents adapt strategies effectively using the proposed crossover.

The method enhances agents' ability to find better strategies over time.

Performance varies under different conditions, showing the method's robustness.

Abstract

We present a simple game which mimics the complex dynamics found in most natural and social systems. Intelligent players modify their strategies periodically, depending on their performances. We propose that the agents use hybridized one-point genetic crossover mechanism,inspired by genetic evolution in biology, to modify the strategies and replace the bad strategies. We study the performances of the agents under different conditions and investigate how they adapt themselves in order to survive or be the best, by finding new strategies using the highly effective mechanism we proposed.