Reactive scheduling approach for solving a realistic flexible job shop scheduling problem

Reactive Scheduling (RS) and the realistic Flexible Job Shop Scheduling Problem (FJSSP) are of major importance for the implementation of real-world manufacturing systems. The present study proposes a scheduling rules-based surrogate assisted simulation-optimisation approach for solving a combinatorial optimisation problem related to a realistic FJSSP. The proposed approach aims to capture the dynamic nature of the FJSSP and to balance both short-term reactivity facing repetitive perturbations and the overall performance of manufacturing systems. Besides and to enhance the optimisation process, a GA-based computational procedure allows managing the use of a hybrid neuronal surrogate and DES model for the accurate and fast calculation of the fitness function, considering the Makespan minimisation criterion and dealing with rush orders. The approach is applied to a highly automated Flexible robotised Manufacturing System (FMS) integrating different realistic and representative constraints to the classical FJSSP. Computational simulations and comparisons demonstrate that the proposed approach shows competitive performances compared to other resolution models, considering obtained solutions quality and short-term reactivity. The proposed resolution model provides technical tools for future control systems and allows for the practical implementation of customised assembly systems in Industry 4.0, relying on innovative emerging technologies.