Scholarlink

<p>Institutions of higher education must balance multiple, often conflicting objectives when setting admission targets for their academic programs. In this paper, we introduce a recommendation system that integrates Constraint Satisfaction Problem (CSP) techniques, goal programming, and Equity Theory to optimize student assignments. Our model strictly enforces hard constraints—such as faculty-hour limits and classroom capacities—while accommodating soft constraints—such as government quotas and institutional preferences—through adjustable penalty functions. Evaluations against static and heuristic benchmarks show that our approach maintains enrollment at 85–90% of total capacity, markedly reducing both the frequency and severity of constraint violations. Furthermore, an average Gini coefficient of 0.067 demonstrates a fairer distribution of seats across programs. Over five simulated admission cycles, institutions employing this recommender achieve substantial compliance improvements within four years, striking an effective balance between rapid constraint adherence and stable enrollment figures. These results confirm that our system offers a practical, data-driven solution for flexible and equitable enrollment management in resource-limited higher-education settings.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

An equity aware recommender system for university admissions balancing operational constraints and strategic objectives

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年11月13日

查看全文
  • Ahmed Ibrahim,
  • Ala Alarood,
  • Eesa Alsolami

摘要

Institutions of higher education must balance multiple, often conflicting objectives when setting admission targets for their academic programs. In this paper, we introduce a recommendation system that integrates Constraint Satisfaction Problem (CSP) techniques, goal programming, and Equity Theory to optimize student assignments. Our model strictly enforces hard constraints—such as faculty-hour limits and classroom capacities—while accommodating soft constraints—such as government quotas and institutional preferences—through adjustable penalty functions. Evaluations against static and heuristic benchmarks show that our approach maintains enrollment at 85–90% of total capacity, markedly reducing both the frequency and severity of constraint violations. Furthermore, an average Gini coefficient of 0.067 demonstrates a fairer distribution of seats across programs. Over five simulated admission cycles, institutions employing this recommender achieve substantial compliance improvements within four years, striking an effective balance between rapid constraint adherence and stable enrollment figures. These results confirm that our system offers a practical, data-driven solution for flexible and equitable enrollment management in resource-limited higher-education settings.