# Merge and Conquer: Evolutionarily Optimizing AI for 2048

arXiv: https://arxiv.org/abs/2510.20205 · HTML: https://mlsanity.com/paper/2510.20205/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 1 of its 6 references that are in the map.
- First posted on arXiv 23 Oct 2025.

## Builds on (most cited first, 1 of 1)

- [Explore the Reasoning Capability of LLMs in the Chess Testbed](https://mlsanity.com/paper/2411.06655/) (2024)

## Abstract (arXiv)

Optimizing artificial intelligence (AI) for dynamic environments remains a fundamental challenge in machine learning research. In this paper, we examine evolutionary training methods for optimizing AI to solve the game 2048, a 2D sliding puzzle. 2048, with its mix of strategic gameplay and stochastic elements, presents an ideal playground for studying decision-making, long-term planning, and dynamic adaptation. We implemented two distinct systems: a two-agent metaprompting system where a "thinker" large language model (LLM) agent refines gameplay strategies for an "executor" LLM agent, and a single-agent system based on refining a value function for a limited Monte Carlo Tree Search. We also experimented with rollback features to avoid performance degradation. Our results demonstrate the potential of evolutionary refinement techniques in improving AI performance in non-deterministic environments. The single-agent system achieved substantial improvements, with an average increase of 473.2 points per cycle, and with clear upward trends (correlation ρ=0.607) across training cycles. The LLM's understanding of the game grew as well, shown in its development of increasingly advanced strategies. Conversely, the two-agent system did not garner much improvement, highlighting the inherent limits of meta-prompting.