A decomposition approach to job-shop scheduling problem with discretely controllable processing times

Ganggang Niu (1, 2) , Shudong Sun (3) , Pascal Lafon (2) , Hongan Yang (3)

1 Institute of System Integration and Engineering Management
2 LASMIS - Laboratoire des Systèmes Mécaniques et d'Ingénierie Simultanée
3 NPU - Northwestern Polytechnical University [Xi'an]

Job-shop scheduling problem with discretely controllable processing times (JSP-DCPT) is modeled based on the disjunctive graph, and the formulation of JSP-DCPT is presented. A three-step decomposition approach is proposed so that JSP-DCPT can be handled by solving a job-shop scheduling problem (JSP) and a series of discrete time-cost tradeoff problems. To simplify the decomposition approach, the time-cost phase plane is introduced to describe tradeoffs of the discrete time-cost tradeoff problem, and an extreme mode-based set dominant theory is elaborated so that an upper bound is determined to cut discrete time-cost tradeoff problems generated by using the proposed decomposition approach. An extreme mode-based set dominant decomposition algorithm (EMSDDA) is then proposed. Experimental simulations for instance JSPDCPT_FT10, which is designed based on a JSP benchmark FT10, demonstrate the effectiveness of the proposed theory and the decomposition approach.

Domaines: Recherche opérationnelle [math.OC]

Ganggang Niu, Shudong Sun, Pascal Lafon, Hongan Yang. A decomposition approach to job-shop scheduling problem with discretely controllable processing times. Science in China Series E: Technological Sciences, 2011, 54 (5), pp.1240-1248. ⟨10.1007/s11431-010-4249-0⟩. ⟨hal-02275891⟩