<![CDATA[Automotive dealers with large sales teams face challenges in fairly distributing customer prospects. Round-robin methods ignore performance differences, while performance-based allocation causes overload and team demotivation. This study develops a Hybrid Weighted Randomization algorithm integrating fairness score (40%), performance score (40%), time decay (20%), and penalty factor. Design Science Research methodology is employed with evaluation through 9 factorial configurations (5, 10, 20 agents; 100, 500, 1000 assignments) measuring Jain's Fairness Index and Coefficient of Variation. The system is implemented using Python 3.13 with PostgreSQL. Results show 88.9% of configurations achieve Grade A+ (JFI ≥ 0.95). Optimal configuration: 10 agents @ 1000 assignments with JFI=0.9916 (0.84% gap from perfect fairness). Three empirical scaling laws are identified: (1) minimum 25 assignments per agent for Grade A+, (2) optimal agent count ≈ √(N_total/10), (3) improvement rate ∝ 1/n_sales². Linear time complexity O(n·m) with 257ms per assignment confirms production-readiness. The algorithm successfully achieves near-optimal fairness while accommodating performance recognition and temporal factors. Scaling laws provide actionable frameworks for team sizing and capacity planning, with theoretical contributions (novel algorithm) and practical ones (production-ready implementation with clear deployment guidelines).]]>
Copyrights © 2026
