Research on the Effect of Teacher Rotation Policy Based on Multi-Agent Simulation: Simulation Evidence from an “Educational Ecosystem” Model

doi:10.16382/j.cnki.1000-5560.2026.08.003

Abstract

Abstract:

This paper is one of the publications featured in the “Human-AI Co-Creation Pioneer Papers Ranking”, which is part of the Panoramic Report on the World’s First Large-Scale Social Experiment of “AI as the First Author” in educational research. Traditional research on education policy often has limitations in evaluating the long-term dynamics and systemic impacts of macro-level policies. To overcome these constraints, this study adopts a computational experiment/simulation research method based on Agent-Based Modeling (ABM) and constructs a computational laboratory named “EduEcosystem”. The model defines three types of heterogeneous agents: students, teachers, and schools, whose attributes and interaction rules are deeply rooted in sociological and psychological mechanisms. Through simulating and deducing the teacher rotation policy, this study reveals a series of complex effects. In the initial stage of policy implementation, it can indeed promote educational equity, with a significant decline in the knowledge Gini coefficient. However, this equity benefit comes at a high cost: the teacher turnover rate rises sharply by about 82.5%, accompanied by a synchronous decline in the overall academic performance level. From the perspective of long-term dynamic evolution, due to the cumulative effect of teacher burnout and adaptive changes within the system, the equity effect of the policy fails to sustain and gradually weakens, eventually giving rise to the counterintuitive phenomenon of “equity rebound”. This study confirms that multi-agent simulation can serve as an effective policy laboratory for understanding complex education systems, and can also preview the long-term dynamic changes and nonlinear chain reactions that policies may trigger. This methodological innovation provides a new approach for the construction of the pedagogical knowledge system and scientific educational decision-making. This study is a computational simulation thought experiment rather than an empirical research, and its conclusions are only for policy reference.

Key words: multi-agent simulation, teacher rotation, educational equity, policy simulation, complex system

DeepSeek, Pei Guo, JingQi Huang, Jin Zhai, Yue Zhou. Research on the Effect of Teacher Rotation Policy Based on Multi-Agent Simulation: Simulation Evidence from an “Educational Ecosystem” Model[J]. Journal of East China Normal University(Educationa, 2026, 44(8): 51-69.

Figures/Tables 11

Table 1

Table 2

Table 3

Table 4

Table 5

Table 6

Table 7

References

	何文萱, 柯政. (2022). 教育政策仿真: 推进教育决策科学化的方法基础. 现代远程教育研究, 34 (04), 65- 72+90.
	胡金艳, 杨霞, 高兴启, 满其峰, 张义兵. (2024). 知识建构社区中多主体模拟支持的观点涌现模型构建: 复杂性科学的视角. 电化教育研究, 45 (01), 100- 107.
	李永梅, 谭维智. (2025). 智能机器驱动的教育学知识生产及其中国化. 现代远程教育研究, 37 (01), 63- 72.
	刘石奇, 刘智, 段会敏, 粟柱, 彭晛. (2025). 大模型驱动的教育多智能体系统应用研究——技术架构、发展现状、实践路径与未来展望. 远程教育杂志, 43 (01), 33- 45.
	苏福根. (2025). 复杂系统视域生成式人工智能促进高等教育发展的风险及治理. 高等工程教育研究, (04), 112- 117.
	孙建文, 冯锦天, 万仟. (2025). 多智能体驱动的虚拟学生模拟助力教师答疑能力提升. 华中师范大学学报(自然科学版), 59 (05), 721- 729.
	王骁汉, 江哲涵, 廖珺, 陈运香, 高燕. (2025). 基于多智能体的PBL模拟教学系统的构建. 护理研究, 39 (13), 2191- 2197.
	谢林, T. C. (2005). 微观动机与宏观行为(谢静, 邓子梁, 李天有, 译). 北京: 中国人民大学出版社.
	杨满福, 桑新民. (2025). 从Deep-Learning到DeepSeek: 人工智能赋能大学功能范式重构的挑战、转型与新生态. 现代教育技术, 35 (04), 5- 13.
	袁磊, 徐济远, 刘沃奇. (2025). 数智教育生态下人机协同教学范式转型. 开放教育研究, 31 (02), 108- 117.
	张佳, 刘智慧, 夏美茹. (2023). 教师交流轮岗政策实施现状及优化对策——利益相关者视角下的混合研究. 教育发展研究, 43 (20), 37- 46.
	Axelrod, R., & Hamilton, W. D. (1981). The evolution of cooperation. Science, 211 (4489), 1390- 1396.
	Arnejo, Z., Gaudou, B., Saqalli, M., & Bantayan, N. (2025). Communicating agent-based models to stakeholders: a scoping review. Journal of Artificial Societies and Social Simulation, 28 (2), 2.
	Belfrage, M., Lorig, F., & Davidsson, P. (2024). Simulating change: a systematic literature review of agent-based models for policy-making. Simulation Series, 56 (1), 512- 524.
	Matthew A. Kraft, William H. Marinell, & Darrick Shen-Wei Yee. (2016). School organizational contexts, teacher turnover, and student achievement: Evidence from panel data. American Educational Research Journal, 53(5), 1411-1449.
	Adnot, M., Dee, T., Katz, V., & Wyckoff, J. (2017). Teacher turnover, teacher quality, and student achievement in DCPS. Educational Evaluation and Policy Analysis, 39(1), 54-76.
	Nguyen, T. D., Pham, L., Crouch, M., & Springer, M. G. (2020). The correlates of teacher turnover: An updated and expanded Meta-analysis of the literature. Educational Research Review, 31, 100355.
	Nespeca, V., Comes, T., & Brazier, F. (2023). A methodology to develop agent-based models for policy support via qualitative inquiry. Journal of Artificial Societies and Social Simulation, 26(1).
	Pensiero, N., & Brede, M. (2024). An agent-based model of school tracking, accountability, and segregation. Journal of Simulation, 18 (4), 524- 538.

属性名称	变量符号	分布类型/初始值	参数依据
基础能力	base_ability	float：N(70,15²)	依据PISA测试成绩分布特征，标准化为均值70，标准差15，以反映学生群体的认知能力差异。
毅力	perseverance	float：Beta(2,2)	依据教育心理学研究，毅力通常不呈正态分布而呈偏态，故采用Beta分布模拟多数学生具有一定毅力、少数学生毅力极强或极弱的现实。
家庭资本	family_capital	float：U(20,80)	依据社会经济地位调查数据，设定其服从均匀分布，以模拟家庭背景的差异性及其对教育的潜在影响。
社交活跃度	social_activity	float：Beta(1.5,1.5)	依据社交行为研究，假设学生的社交倾向呈对称分布，采用对称参数Beta分布进行模拟。
知识水平	knowledge	float	动态状态变量，表示学生在时间步t所掌握的知识总量，初始值由其基础能力和家庭资本共同决定。
学习容量	learning_capacity	float	受基础学业能力和家庭文化资本影响的个体学习上限，体现“天花板效应”，其计算公式见下文。

模块	数学表达/伪代码	说明
知识增长模块	$ \begin{array}{l}learning\text{}\_\text{}capacit{y}_{i}=50+0.3 \times base\text{}\_\text{}abilit{y}_{i}\\\qquad\qquad\qquad\qquad +0.2 \times family\text{}\_\text{}capita{l}_{i}\end{array} $	学习容量计算：综合先天能力与后天环境设定个体成长极限。
知识增长模块	$ \begin{array}{l}grow\text{}\_\text{}rate=0.08 \times T{E}_{teac\text{h}er}(t) \times \left(1+\dfrac{S{C}_{sc\text{h}ool}}{100}\right) \\\qquad\qquad\qquad \times \left(0.6+0.4 \times perseveranc{e}_{i}\right) \end{array}$ $ \begin{array}{l}{K}_{i}(t+1)={K}_{i}(t)+grow\text{}\_\text{}rate \times {K}_{i}(t) \times \left(1-\dfrac{{K}_{i}(t)}{learning\text{}\_\text{}capacit{y}_{i}}\right)\\\qquad\qquad\qquad +N(0,0.{3}^{2}) \end{array}$	逻辑斯蒂增长模型：此为核心创新点，模拟知识增长的非线性过程。初始阶段快速提升，接近容量极限时增速放缓，有效避免传统线性模型的失真，更贴合真实学习规律。
社交影响模块	if random()<social_activity_i: j=随机选择社交网络邻居 if knowledge_j>knowledge_i: # 非线性收益：差距适中时收益最大 $ gap=knowledg{e}_{j}-knowledg{e}_{i} $ optimal_gap = 15 # 最佳学习差距 $ \begin{array}{l}learning\_ gain= 0.04 \times social\_ activity\_ i\times gap\times exp(-(gap\\\qquad\qquad\qquad -optimal\_ gap){^{2}}/(2\times optimal\_ gap{^{2}})) \end{array}$$ knowledg{e}_{i}+=learning\_ gain $	非线性社交学习：此机制体现经典的“最近发展区”理论。当学生i与同伴j的知识水平存在适度差距（设定最佳差距optimal_gap为15）时，学习收益最大；差距过小则缺乏学习动力，差距过大则超出理解范围。这精准地模拟同辈效应的复杂性与条件性。

属性名称	变量符号	分布类型/初始值	参数依据
教学效能	teaching_efficacy	float：N(75, 15²)	依据教师效能研究的分布特征，设定其服从正态分布，反映教师队伍教学能力的自然差异。
毅力	adaptability	float：Beta(2.5,2)	依据组织行为学理论，教师适应新环境的能力呈偏态分布，故采用Beta分布模拟多数教师具备中等以上适应能力的现实。
家庭背景	family_capital	float：U(20,80)	依据社会经济地位调查数据，设定其服从均匀分布，以模拟家庭背景的差异性及其对教育的潜在影响。
抗压能力	stress_resilience	float：Beta(2, 3)	依据职业心理学研究，个体对工作压力的复原力存在差异，采用Beta分布模拟其异质性。
恢复计数器	recovery_counter	Int: 0	状态变量，记录教师轮岗后经历的恢复时间，用于计算其教学效能的恢复程度。

模块	数学表达/伪代码	说明
职业倦怠动态（非线性累积）	teaching_pressure = (班级能力差异系数) × (班级规模/35)	工作压力源
	administrative_pressure = 0.3	行政事务基础压力
	adjustment_pressure = 0.0	轮岗适应压力
	adjustment_pressure = 0.4 × (1 − adaptability)	轮岗状态
	$ pressur{e}_{total}=teac\text{h}ing\text{}\_\text{}pressure+administrative\text{}\_\text{}pressure+ $$ adjustment\text{}\_\text{}pressure $$ \begin{array}{l}burnout\_ increment=0.025\times pressure\_ total\times (1 \\\qquad\qquad\qquad + 0.5 \times burnout{^{2}})\times (1 - 0.3 \times stress\_ resilience) \end{array}$$ burnout = min(1.0, burnout + burnout\_ increment) $	非线性倦怠累积模型
	if pressure_total < 0.4: $ burnout=\max \left(\begin{array}{c}0.1,burnout-0.015\times (1-burnout)\times \\stress\_ resilienc{e}_{i}\end{array}\right) $	恢复机制（非线性衰减）
教学效能计算	$ efficacy\_ reduction = 0.5 \times (1 - exp(-2 \times burnout)) $$\begin{array}{l} effective\_ efficacy = teaching\_ efficacy \times (1 \\ \qquad\qquad\qquad- efficacy\_ reduction) \end{array}$	效能受倦怠影响的非线性衰减：采用指数函数模拟倦怠对效能的侵蚀作用，初期影响显著，后期趋于平缓，符合边际效用递减规律。
教学效能计算	if 轮岗状态 and recovery_counter < 恢复周期: $\begin{array}{l} adaptatio\_ penalty=0.4 \times (1-adaptability)\\ \qquad\qquad \times exp(-recovery\_ counter/5) \end{array}$$ effective\_ efficacy = (1 - adaptation\_ penalty) $	轮岗适应期调整：轮岗初期，教师的教学效能会因其适应能力不足而受到额外惩罚，该惩罚随恢复时间的推移呈指数衰减，模拟教师在新环境中的逐步适应过程。

属性名称	变量符号	计算与更新方式	参数与理论依据
平均文化资本	avg_family_capital	每学期初计算	反映学校的生源社会经济背景，是衡量校际差异和教育起点的关键指标。
组织氛围	organizational_climate	动态合成计算，详见下方规则。	综合反映学校的教学环境质量，是影响师生效能的关键组织变量。
教师凝聚力	teacher_cohesion	（可选）可基于教师流失率与校龄结构反向估算。	表征教师队伍的稳定性与协作程度，是组织氛围的组成部分。