A Local Search-Based Generalized Normal Distribution Algorithm for Permutation Flow Shop Scheduling

This paper studies the generalized normal distribution algorithm (GNDO) performance for tackling permutation flow shop scheduling problem (PFSSP). Because PFSSP is a discrete and GNDO generates continuous values, largest ranked value rule used to convert those values into ones make applicable solving this problem. Additionally, effectively integrated with local search strategy improve quality of best-so-far solution in an abbreviated version HGNDO. More than that, new improvement using swap mutation operator applied on avoid being stuck optima by accelerating convergence speed HGNDO propose version, namely hybrid-improved (HIGNDO). Last but not least, improved scramble utilize each trial as ideally possible reaching better outcomes. IGNDO produce strong IHGNDO. Those proposed algorithms are extensively compared number well-established optimization various statistical analyses estimate optimal makespan 41 well-known instances reasonable time. The findings show benefits speedup both IHGNDO HIGNDO over all algorithms, addition