Optimizing headcount decisions in the apparel industry: Leveraging Predictive Analytics and Machine Learning

Abstract

ABSTRACT Labour management, especially in labour-intensive sectors like apparel manufacturing, is crucial for maintaining operational efficiency and managing costs effectively. The COVID-19 pandemic and ensuing economic challenges have worsened these concerns, prompting the need for innovative solutions. In light of devising a solid mechanism to facilitate headcount-related decisions, this research focuses on Emjay Penguin, a prominent garment manufacturer in Sri Lanka, which sought to optimize its warehouse department's headcount management amidst changing demands and digital transformation initiatives. A data-driven approach was adopted, leveraging machine learning algorithms to predict optimal worker and staff cadre requirements. Through rigorous experimentation and performance evaluation, Gradient Boosting Regressor emerged as the most effective model, which together with the Suffering Index offering precise predictions and insights into headcount reduction or reallocation potential. Key features influencing cadre prediction were identified, aiding informed decision-making. The research successfully achieved its objectives, providing a robust framework for headcount management and cost reduction. The findings underscore the importance of aligning labour resources with production targets and demonstrate the value of advanced technologies in enhancing operational efficiency. Keywords: Labour management, apparel industry, headcount optimization, machine learning, Gradient Boosting Regressor, cost reduction, data-driven decision-making.