Bridging Multi-Valued Heuristics and Dimensionality Reduction in Multi-Objective Search

The article presents a new theoretical framework for integrating multi-valued heuristics (MVHs) with dimensionality reduction (DR) in multi-objective shortest-path (MOSP) algorithms. It introduces the $\text{L}\text{-}\text{NAMOA}^*{\text{dr}\text{-}\text{mvh}}$ algorithm, which maintains search correctness by dynamically addressing local ordering violations while utilizing an admissible MVH. This approach demonstrates significant performance improvements, achieving speedups of over 10x compared to existing state-of-the-art MOSP algorithms, making it a valuable advancement for practitioners seeking efficient multi-objective search solutions.