.png){kind=icon}
.png){kind=icon}
Return
Intelligent resource planning optimization: enhancing decision making with artificial intelligence
.png){kind=icon}
PRE
AI
DOI：10.1007/s12351-025-01002-3.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
A professional service organization (PSO) must allocate workforce resources to project jobs, a process known as resource planning (RP). We propose an Intelligent Resource Planning Optimization (IRPO) framework that integrates natural language processing (NLP) with optimization techniques to enhance this task. Using a fine-tuned Bidirectional Encoder Representations from Transformers (BERT) classifier, we extract job attributes (skills, education, experience) from curriculum vitae (CVs) and job postings and compute resource-job matching scores via Sentence Transformers (ST) embeddings. To ensure high-quality matching, we compute a resource-job score that considers multiple job attributes and hiring manager priorities. These scores feed into a reformulated multi-period resource planning problem, expressed as a minimum-cost matching (MCM) model on a bipartite graph and solved to optimality with Bertsekas' \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\varepsilon $$\end{document}-scaling auction algorithm. Computational experiments with real CVs and job postings demonstrate that our custom NLP models achieve a CV ranking closely aligned with expert judgment, that IRPO is orders of magnitude faster than a Gurobi-based mixed-integer linear programming (MILP) approach, and that it achieves 99.7% demand fulfillment and 84% capacity utilization-outperforming greedy heuristics by over 20%. For practitioners, IRPO offers a pathway to superior workforce management by integrating assignment, training, and hiring decisions into a single optimization. This holistic approach directly translates to enhanced demand fulfillment and resource utilization, reducing project delays and staffing costs. By providing attribute-based, explainable scores, our framework also fosters trust in artificial intelligence (AI)-driven decision support and presents a scalable blueprint for workforce planning in other complex sectors.
Keywords:
Multiperiod resource planning
Natural language processing
Assignment problem
Professional services organizations
Journal
IF:
2.7
Papers:
77
Citations:
1.7K
View more organizations