Akademik Veri Yönetim Sistemi
NEURAL COMPUTING & APPLICATIONS, cilt.33, sa.8, ss.3469-3486, 2021 (SCI-Expanded)
The permutation flow shop scheduling problem has been investigated by researchers for more than 40 years due to its complexity and lots of real-life examples of the problem. Many exact or approximate solution approaches have been presented for the problem. Among solution approaches in the literature, iterated greedy algorithm and its variants are the most effective solution methods for the problem. This paper proposes a hybrid solution algorithm that uses the best components such as local search operators and destruction/construction operators of the variants of iterated greedy algorithm in an artificial bee colony algorithm. An ANOVA is made for determining the most proper components of iterated greedy algorithm. Then, these components are combined with artificial bee colony algorithm. Furthermore, a design of experiment is made for determining the best parameter setting for the proposed hybrid artificial bee colony. The experimental results of the proposed algorithm compared with variants of iterated greedy algorithms in the literature show that the proposed algorithm produces better solutions that deviate less from the optimum or lower-bound solutions for permutation flow shop scheduling problems with the makespan performance criterion.