Iterated greedy algorithms for a complex parallel machine scheduling problem

This paper addresses a complex parallel machine scheduling problem with jobs divided into operations and grouped in families. Non-anticipatory family setup times are held at the beginning of each batch, defined by combination one setup-time sequence from unique family. Other aspects also considered problem, such as release dates for machines, operation’s sizes, machine’s eligibility capacity. We consider item availability to define completion within batches, minimize total weighted time jobs. developed Iterated Greedy (IG) algorithms combining destroy repair operators Random Variable Neighborhood Descent (RVND) local search procedure, using four neighborhood structures solve problem. The best algorithm variant outperforms current literature methods terms average deviation solutions computational times, known benchmark set 72 instances. New upper bounds provided some instances this set. Besides, experiments conducted evaluate proposed methods’ performance more challenging introduced work. Two IG variants greedy operator showed superior than 70% found uniquely these variants. Despite simplicity, method most common destruction presented results different evaluated criteria, highlighting its potential applicability solving