Adaptive-resolution simulation: allocate more LLM calls to educationally consequential moments and compress routine transitions¶

Strategy · All strategies
Evidence · no claims cited

Description¶

To manage the coupled trade-off among scale, duration, and granularity, the paper recommends making the simulation budget configurable and concentrating generative detail where it matters. As printed: "Efficiency is achieved by adaptive resolution. The simulator can allocate more LLM calls to educationally consequential moments, such as misconception diagnosis, teacher reflection, conflict resolution, or high-stakes assessment, while compressing routine transitions into structured updates." Caching of stable agent profiles and memory retrieval further reduce runtime while preserving validation-relevant detail.

Design Implications¶

Context¶

Requirements¶

• Mixed granularity support so important lessons can be simulated turn by turn while routine practice is compressed into state updates with summary evidence

Constraints¶

• The simulation budget B ≈ N·H·(1/Δt)·c_step is finite, so every simulator implicitly decides which educational mechanisms remain visible

Target Learners¶

• Researchers running multi-agent educational simulations across turn, lesson, and semester granularities

Target Learning Goals¶

• Preserving visibility of misconception repair, scaffolding, and social dynamics within computational budgets

Affordances¶

• Education As Multi Agent State Transition

Related Strategies¶

• Time Management In Instruction
• Team training with variable leadership roles and simulated communication delays
• Design hybrid instructional chatbot systems that balance breadth and depth via multi-agent architectures or adaptive interaction modes

Examples¶

-

Key Sources¶

• Yulei Ye, Wenhao Li, Zhong Wen, Yunshu Huang, Yichen Hu, Zifan Wei, Yige Wang, Xinyu Xie, Haoxuan Yang, Yanjun Huang, Ruijia Li, Hong Qian, Yu Song, Bo Jiang, Bingdong Li, Lijun Li, Bo Zhang, Pinlong Cai, Xingcheng Xu, Shuangye Chen, Xia Hu, Liang He, Aimin Zhou, Jingjing Qu, Jing Shao, Xiangfeng Wang. (2026). AgentSchool: An LLM-Powered Multi-Agent Simulation for Education. arXiv. https://arxiv.org/abs/2605.30144