Optimizing Work Schedule Assignments for Straddle Carrier Drivers at Container Terminals: A Collaborative Filtering Recommender System Approach

This paper introduces a novel collaborative filtering recommender system designed to optimize work schedule assignments for Straddle Carrier (SC) drivers at container terminals. The proposed Straddle Carrier Assignment Model (SAM) addresses critical operational challenges by integrating multi-dimensional rating matrices with seniority-based similarity metrics to create an intelligent scheduling system that balances operational efficiency with workforce satisfaction.

The system was implemented at the RADES container terminal using a three-tier architecture that incorporates real-time feedback mechanisms and an intelligent scoring algorithm that dynamically adapts to changing operational conditions. The mathematical framework combines collaborative filtering with domain-specific constraints through hybrid similarity computation, dynamic neighbor selection, and constrained optimization algorithms.

The implementation demonstrated significant operational improvements, including a 93% reduction in schedule response time, a 64% decrease in assignment disputes, and a 31% increase in container handling efficiency, over a 24-month evaluation period. The system achieved 99.9% uptime, with a 28% improvement in resource utilization and an 85% positive driver satisfaction rating.

SAM's innovative approach represents a significant advancement over traditional rule-based scheduling methods by introducing machine learning techniques to the maritime logistics domain. The mathematical framework combines collaborative filtering with domain-specific constraints to produce schedules that optimize both terminal productivity and driver satisfaction.

By addressing the fundamental challenges of schedule optimization in container terminals, this research provides both theoretical contributions to recommender systems and practical value to maritime logistics operations.