Evaluation of Decision Optimization Model using Deep Reinforcement Learning in Collaborative School-Enterprise-Local Education Decision Making

Abstract

The school-enterprise model fosters cooperation between academic institutions, local businesses, and communities to enhance students' hands-on learning experiences. This study investigates the application of Deep Reinforcement Learning (DRL) in analyzing the mechanisms and current status of school-enterprise collaboration. We propose a novel Distance-Driven Proximal Policy Optimization (DD-PPO) framework to optimize this cooperation. Case studies and assessments were conducted in Hangzhou City's higher vocational institutions to evaluate the model's effectiveness. Findings reveal that 55% of businesses prioritize recruiting fresh graduates, while half of the schools perceive the collaboration as universally beneficial. The professional orientation of schools, driven by business demands, enhances graduate employability but lacks systematic planning, policy support, and alignment in evaluation metrics. These disparities complicate the establishment of a seamless and efficient school-enterprise linkage. Additionally, the inconsistencies in practical skill development among students from different programs highlight the need for a more structured and adaptable system. The proposed DD-PPO model addresses these challenges, offering a promising solution for enhancing school-enterprise partnerships.