Optimization of Permutation Flowshop Schedulling Problem (PFSP) using First Sequence Artificial Bee Colony (FSABC) Algorithm

Abstract

The most widely used production system in the sector is the flowshop, and several documented initiatives have been made to enhance its functionality. The scope extends from heuristic application to metaheuristics; NEH (Nawaz, Enscore & Ham) heuristics are among the best techniques. This work attempts to improve NEH by utilising an improved version of the slow-to-converge Artificial Bee Colony (ABC) method. This study will thus suggest a method to increase ABC's convergence speed since, in comparison to prior studies, a quicker convergence speed allows for the discovery of superior findings in a shorter amount of time. Because it produced the best makespan, the Total Greedy (5+0+0 & 10+0+0) was selected for Employed Bees (EB) and Onlooker Bees (OB), and the author employed it for the duration of the study. This study suggested using the First Job Sequence Arrangement Method into Artificial Bee Colony thus resulting in First Sequence Artificial Bee Colony, or FSABC. Two Taillard benchmark datasets were used for the study's trials, which were split into two groups: 20 jobs 5 machines (5EBOB) and 50 jobs 5 machines (10EB10OB). Each dataset is run through numerous iterations of the experiments 500, 1000, and 2000 iterations total with 20 replications. This study examined the effectiveness of NEH, ABC, and FSABC, which also acts as the validation process. The study reveals that while FSABC may achieve the required results with a very good improvement over the original ABC and NEH, ABC regularly produces inconsistent results, and the majority of its data is inferior to NEH. The author can now say with certainty that the slow convergence problem of ABC has been resolved by this study.