File(s) under permanent embargo
Multi-objective job shop scheduling using i-NSGA-III
conference contribution
posted on 2018-01-01, 00:00 authored by Burhan KhanBurhan Khan, Samer HanounSamer Hanoun, Michael JohnstoneMichael Johnstone, Chee Peng LimChee Peng Lim, Douglas CreightonDouglas Creighton, Saeid NahavandiThe complexity of job shop scheduling problems is related to many factors, such as a large number of jobs, the number of objectives and constraints. Evolutionary algorithms are a natural fit to search for the optimum schedules in complex job shop scheduling problems with multiple objectives. This paper extends the authors' i-NSGA-In algorithm to tackle a manufacturing job shop scheduling problem with multiple objectives. One of the complex objectives is to pair jobs with similar properties to increase the overall cost savings. The genetic operators in i-NSGA-III are replaced with novel problem-specific crossover and mutation operators. The proposed approach is validated by comparing against the enumeration technique for problems with 5 to 10 jobs. Unlike the enumeration technique, the proposed methodology shows competence in terms of computation time and ability to schedule a large number of jobs with a high number of objectives. Further comparisons with NSGA-III demonstrate the superiority of i-NSGA-III for problems with 30, 40, and 50 jobs.