An experimental evaluation of interval-based adaptations of the Jonker–Volgenant algorithm
This work investigates the impact of uncertainty in the Linear Assignment Problem through an experimental comparison between the classical Jonker–Volgenant algorithm and an interval-based extension. While the classical approach assumes precise cost values, many real-world applications involve imprecise or variable data that cannot be adequately represented by single numerical values.
The study evaluates both methods under different uncertainty scenarios, including deterministic, low-variability, and high-variability settings, focusing on robustness, stability, and solution quality. The results indicate that the use of interval costs provides a more expressive representation of uncertainty, leading to more robust and informative assignment solutions.
These findings suggest that interval-based optimization constitutes a valuable alternative to deterministic approaches, particularly in decision-making contexts where data uncertainty plays a significant role.
Keywords: Linear Assignment Problem Jonker–Volgenant algorithm interval data uncertainty optimization decision-making