Genetic programming hyper-heuristic-based solution for dynamic energy-efficient scheduling of hybrid flow shop scheduling with machine breakdowns and random job arrivals

Abstract

Aiming at the lack of scientific methods for solving the dynamic energy-efficient scheduling problem of hybrid flow shop using scheduling rules, this paper proposes a method that can automatically generate scheduling rules based on the processing information of the shop. Firstly, a multi-objective mathematical model with the objective of minimizing the maximum tardiness, machine idle energy consumption and maximum makespan is established by combining two dynamic events, namely, machine breakdowns and random job arrivals. Secondly, a genetic programming hyper-heuristic algorithm, utilizing terminal sets to generate high-level scheduling rules, is employed for the dynamic energy-efficient hybrid flow shop scheduling problem. Considering dynamic energy-efficient scheduling of the shop, terminal sets for two dynamic events and energy-efficient objects are designed, and the performance of the scheduling rules is improved by assigning weight coefficients to each terminal. Finally, comparisons of the scheduling rules generated by the proposed method and the benchmark scheduling rules are conducted in 36 scenarios. The result demonstrate that the algorithm has a high degree of flexibility and adaptability.