Multi-objective biased randomised iterated greedy for robust permutation flow shop scheduling problem under disturbances

Nowadays, scheduling problems under different disruptions are a key to become competitive in the global market of this century. Permutation flow shop scheduling problems are very important as they consider one of the important types of scheduling problems. In this paper, we consider a challenging scheduling problem of a permutation flow shop in the presence of different types of real-time events such as new job arrival and machine breakdown. A multi-objective optimisation model that takes into account multiple performance measures in order to minimise the effect of different real-time events is used in this paper. To solve this problem, we apply the proposed multi-objective model and adapt a predictive-reactive based Biased Randomised Iterated Greedy approach for the problem, which is hybridised a Biased Randomisation process and the Iterated Greedy algorithm. Furthermore, the proposed approach is compared against the predictive-reactive based Particle Swarm Optimisation method for the same problem. Additionally, to show the efficiency of the proposed model, we compare this model by testing the predictive-reactive based BRIG approach to two other models: the bi-objective model that consider only two objectives and the classical single-objective model of minimising the makespan. Further statistical analysis is performed in this study by using an Analysis of Variance measure. The extensive experiments and statistical analysis demonstrate that the proposed multi-objective model is better than the other models in reducing the relative percentage deviation. Additionally, despite their simplicity, the BRIG algorithm is shown to be state-of-the-art method that outperforms the Particle Swarm Optimisation algorithm.