A multi-objective invasive weed optimization algorithm for robust aggregate production planning under uncertain seasonal demand

• Published in: Computing Vol. 101; no. 6; pp. 499 - 529
• Main Authors: Goli, Alireza, Tirkolaee, Erfan Babaee, Malmir, Behnam, Bian, Gui-Bin, Sangaiah, Arun Kumar
• Format: Journal Article
• Language: English
• Published: Vienna Springer Vienna 01.06.2019; Springer Nature B.V
• Subjects: Artificial Intelligence; Classification; Computer Appl. in Administrative Data Processing; Computer Communication Networks; Computer Science; Demand; Employment; Genetic algorithms; Goal programming; Information Systems Applications (incl.Internet); Integer programming; Linear programming; Mathematical models; Mixed integer; Multiple objective analysis; Optimization; Optimization algorithms; Production planning; Production scheduling; Robustness (mathematics); Software Engineering; Sorting algorithms; Taguchi methods; NSGA-II; 68Txx; 90B30; Robust optimization; Aggregate production planning; 90C59; Uncertain seasonal demand; Multi-objective invasive weed optimization algorithm (MOIWO); 90B50
• DOI: 10.1007/s00607-018-00692-2
• ISSN: 0010-485X

This paper addresses a robust multi-objective multi-period aggregate production planning (APP) problem based on different scenarios under uncertain seasonal demand. The main goals are to minimize the total cost including in-house production, outsourcing, workforce, holding, shortage and employment/unemployment costs, and maximize the customers’ satisfaction level. To deal with demand uncertainty, robust optimization approach is applied to the proposed mixed integer linear programming model. A goal programming method is then implemented to cope with the multi-objectiveness and validate the suggested robust model. Since APP problems are classified as NP-hard, two solution methods of non-dominated sorting genetic algorithm II (NSGA-II) and multi-objective invasive weed optimization algorithm (MOIWO) are designed to solve the problem. Moreover, Taguchi design method is implemented to increase the efficiency of the algorithms by adjusting the algorithms’ parameters optimally. Finally, several numerical test problems are generated in different sizes to evaluate the performance of the algorithms. The results obtained from different comparison criteria demonstrate the high quality of the proposed solution methods in terms of speed and accuracy in finding optimal solutions.