Two-stage dynamic scheduling model for a flexible flow shop using a genetic algorithm: a case study in a truck body manufacturing company

Tóm tắt

Dynamic flexible flow shop scheduling problems present a complex challenge, where new orders arrive stochastically during the current schedule; this uncertainty degrades schedule stability and resource utilization, highlighting the need for responsive scheduling mechanisms. This study proposes a two-stage scheduling framework. The “master schedule” drives global coordination across all stages, while the “reactive” schedule enables immediate rescheduling decisions in response to uncertain job arrivals within the current “master schedule” through the use of a dispatching rule. The problem is modeled using mixed-integer programming (MIP). The objective is to minimize the makespan, and the problem is solved by a genetic algorithm (GA). Experimental results demonstrate that the proposed model outperforms traditional dispatching rules. It achieves a makespan reduction of up to 4.9%, a tardiness reduction of over 40%, and a flow time improvement of over 24%. The study provides a practical and scalable solution for dynamic flexible flow shop environments.