Resolving the Prisoner’s Dilemma in Industry, University, and Research Cooperation: A Three Party Evolutionary Game with BERT Based Parameter Estimation
- 1College of Mathematics and Computer Science, Guangdong Ocean University, Zhanjiang 524088, China
Abstract
Under the Industry-University-Research (IUR) cooperation framework, universities, enterprises and college students, driven by short-term interests, are prone to fall into the “Prisoner’s Dilemma”, which restricts the depth, breadth and sustainability of the cooperation. This study establishes a tripartite evolutionary game model that treats universities, enterprises, and students as equal decision makers, and calibrates its parameters empirically rather than by assumption: a BERT-based semantic model converts the text records of 156 real IUR projects (Guangdong, 2020–2023) into game parameters, which are cross-checked against a questionnaire survey of 314 students (287 valid responses, 91.4% validity). MATLAB numerical simulations show that the system converges to a unique globally stable cooperative equilibrium at (1,1,1), in which all three parties cooperate, once the three eigenvalue stability conditions hold under the empirically calibrated parameters; these conditions include the university factor T + αW > 2H + BU + λL. The main quantified results are: raising the incentive fund H from 12.5 to 17.5 accelerates early convergence by about 35%; penalties on enterprises and students reduce their non-cooperation by 42.3% and 37.8%, respectively; the enterprise benefit share contributes roughly 2.3 times as much to system stability as the student share; and a passive university drags all partners into cascading losses. The contributions include a tripartite game formulation that endogenizes students as strategic actors, a transparent sentiment-to-parameter pipeline, and empirically calibrated stability thresholds. These findings provide empirically grounded benchmarks for setting incentives, penalties, and supervision, and show that well-designed institutional rules can resolve the dilemma through the system’s own self-reinforcing dynamics, reducing the need for constant external intervention.
Related
- Diffusion-induced instabilities promote cooperation in eco-evolutionary networks
- Parameter Estimation for Eccentric Supermassive Black Hole Binaries with Pulsar Timing Arrays
- Selecting movements for muscle parameter estimation: insights from gradient analysis in trajectory optimization
- Hybrid PINN-TWM framework for DFB laser modeling and parameter estimation under small-sample constraints
- Optimal Hamiltonian Parameter Estimation in the Presence of Nuisance Parameters
BibTeX
@article{chen2026resolving,
title = {{Resolving the Prisoner’s Dilemma in Industry, University, and Research Cooperation: A Three Party Evolutionary Game with BERT Based Parameter Estimation}},
author = {Rongning Chen and Jiuji Zhang and Yifei Hua and Jieqiong Xiao},
journal = {Journal of Applied Science and Engineering},
year = {2026},
month = oct,
doi = {10.6180/jase.202701_36.043},
url = {https://doi.org/10.6180/jase.202701_36.043},
note = {CC BY},
}
RIS
TY - JOUR
TI - Resolving the Prisoner’s Dilemma in Industry, University, and Research Cooperation: A Three Party Evolutionary Game with BERT Based Parameter Estimation
AU - Rongning Chen
AU - Jiuji Zhang
AU - Yifei Hua
AU - Jieqiong Xiao
PY - 2026
DA - 2026/10/31
JO - Journal of Applied Science and Engineering
DO - 10.6180/jase.202701_36.043
UR - https://doi.org/10.6180/jase.202701_36.043
AB - Under the Industry-University-Research (IUR) cooperation framework, universities, enterprises and college students, driven by short-term interests, are prone to fall into the “Prisoner’s Dilemma”, which restricts the depth, breadth and sustainability of the cooperation. This study establishes a tripartite evolutionary game model that treats universities, enterprises, and students as equal decision makers, and calibrates its parameters empirically rather than by assumption: a BERT-based semantic model converts the text records of 156 real IUR projects (Guangdong, 2020–2023) into game parameters, which are cross-checked against a questionnaire survey of 314 students (287 valid responses, 91.4% validity). MATLAB numerical simulations show that the system converges to a unique globally stable cooperative equilibrium at (1,1,1), in which all three parties cooperate, once the three eigenvalue stability conditions hold under the empirically calibrated parameters; these conditions include the university factor T + αW > 2H + BU + λL. The main quantified results are: raising the incentive fund H from 12.5 to 17.5 accelerates early convergence by about 35%; penalties on enterprises and students reduce their non-cooperation by 42.3% and 37.8%, respectively; the enterprise benefit share contributes roughly 2.3 times as much to system stability as the student share; and a passive university drags all partners into cascading losses. The contributions include a tripartite game formulation that endogenizes students as strategic actors, a transparent sentiment-to-parameter pipeline, and empirically calibrated stability thresholds. These findings provide empirically grounded benchmarks for setting incentives, penalties, and supervision, and show that well-designed institutional rules can resolve the dilemma through the system’s own self-reinforcing dynamics, reducing the need for constant external intervention.
SN - 2708-9967
SN - 2708-9975
ER -