Research on Repair Process Scheduling Algorithm Based on Reinforcement Point Network

Le résumé fourni par la source

Repair process scheduling represents a typical combinatorial optimization problem, where rational sequencing of repair operations can significantly reduce the overall maintenance duration and enhance repair efficiency. To address the limitations of conventional algorithms in terms of poor generalization capability and low computational efficiency, this paper proposes a novel deep reinforcement learning-based approach for solving repair process scheduling problems with superior generalizability. Experimental results on the Ta benchmark dataset demonstrate that our algorithm achieves an average relative error of 4.22%, outperforming CDS, Palmer, CBP, and GDAL algorithms by reducing their average relative errors by 2.62%, 2.00%, 5.48%, and 0.17%, respectively. Moreover, the proposed model exhibits remarkable computational efficiency, requiring only 0.13 seconds on average to solve Ta benchmark instances.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.