# Self-Evolving Multi-Agent Framework for Efficient Decision Making in Real-Time Strategy Scenarios Source: [Self-Evolving Multi-Agent Framework for Efficient Decision Making in Real-Time Strategy Scenarios]() Collection index: [llms.txt]() Data updated: 2026-03-27T05:00:04.194Z Authors: Li Ma, Hao Peng, Yiming Wang, Hongbin Luo, Jie Liu, Kongjing Gu, Guanlin Wu, Hui Lin, Lei Ren Published: 2026-03-25T03:05:29.000Z ## Abstract Large language models (LLMs) have demonstrated exceptional potential in complex reasoning,pioneering a new paradigm for autonomous agent decision making in dynamic settings. However, in Real-Time Strategy (RTS) scenarios, LLMs suffer from a critical speed-quality trade-off. Specifically expansive state spaces and time limits render inference delays prohibitive, while stochastic planning errors undermine logical consistency. To address these challenges, we present SEMA (Self-Evolving Multi-Agent), a novel framework designed for high-performance, low-latency decision-making in RTS environments. This collaborative multi-agent framework facilitates self-evolution by adaptively calibrating model bias through in-episode assessment and cross-episode analysis. We further incorporate dynamic observation pruning based on structural entropy to model game states topologically. By distilling high dimensional data into core semantic information, this approach significantly reduces inference time. We also develop a hybrid knowledge-memory mechanism that integrates micro-trajectories, macro-experience, and hierarchical domain knowledge, thereby enhancing both strategic adaptability and decision consistency. Experiments across multiple StarCraft II maps demonstrate that SEMA achieves superior win rates while reducing average decision latency by over 50%, validating its efficiency and robustness in complex RTS scenarios. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Global resource management and long-term strategic planning on Melee Flat32 Lv.1 (Very Easy)](): Global resource management and long-term strategic planning; Melee Flat32 Lv.1 (Very Easy); Win Rate: 100. - [Global resource management and long-term strategic planning on Melee Flat32 Easy Lv.2](): Global resource management and long-term strategic planning; Melee Flat32 Easy Lv.2; Win Rate: 94. - [Global resource management and long-term strategic planning on Melee Flat48 Lv.2 (Easy)](): Global resource management and long-term strategic planning; Melee Flat48 Lv.2 (Easy); Win Rate: 95. - [Global resource management and long-term strategic planning on Melee Flat64 Lv.1 (Very Easy)](): Global resource management and long-term strategic planning; Melee Flat64 Lv.1 (Very Easy); Win Rate: 100. - [Global resource management and long-term strategic planning on Melee Flat64 Easy Lv.2](): Global resource management and long-term strategic planning; Melee Flat64 Easy Lv.2; Win Rate: 94. - [Global resource management and long-term strategic planning on Melee Simple64 Lv.1 (Very Easy)](): Global resource management and long-term strategic planning; Melee Simple64 Lv.1 (Very Easy); Win Rate: 100. - [Global resource management and long-term strategic planning on Melee Simple64 Lv.2 (Easy)](): Global resource management and long-term strategic planning; Melee Simple64 Lv.2 (Easy); Win Rate: 100. - [Multi-agent Micromanagement on SMAC 3m](): Real-time micro-management; SMAC 3m; Win Rate: 88. - [Real-time micro-management on SMAC 8m](): Real-time micro-management; SMAC 8m; Win Rate: 70. - [Real-time micro-management on SMAC 25m](): Real-time micro-management; SMAC 25m; Win Rate: 68. - [Global resource management and long-term strategic planning on Melee Flat48 Lv.1 (Very Easy)](): Global resource management and long-term strategic planning; Melee Flat48 Lv.1 (Very Easy); Win Rate: 1. ## Original sources and implementation [Original paper]() [PDF]()