A new multiobjective tiki-taka algorithm for optimization of assembly line balancing

• Main Authors: M. F. F., Ab Rashid, Ariff Nijay, Ramli
• Format: Article
• Language: English; English
• Published: Emerald Group Publishing Ltd. 2023
• Subjects: TJ Mechanical engineering and machinery; TS Manufactures
• Online Access: http://umpir.ump.edu.my/id/eprint/37577/1/2023%20MOTTA%20Eng%20Comp.pdf; http://umpir.ump.edu.my/id/eprint/37577/7/A%20new%20multiobjective%20tiki-taka%20algorithm%20for%20optimization%20of%20assembly%20line%20balancing.pdf; http://umpir.ump.edu.my/id/eprint/37577/; https://doi.org/10.1108/EC-03-2022-0185; https://doi.org/10.1108/EC-03-2022-0185
• Institution: Universiti Malaysia Pahang

Purpose: This study aims to propose a new multiobjective optimization metaheuristic based on the tiki-taka algorithm (TTA). The proposed multiobjective TTA (MOTTA) was implemented for a simple assembly line balancing type E (SALB-E), which aimed to minimize the cycle time and workstation number simultaneously. Design/methodology/approach: TTA is a new metaheuristic inspired by the tiki-taka playing style in a football match. The TTA is previously designed for a single-objective optimization, but this study extends TTA into a multiobjective optimization. The MOTTA mimics the short passing and player movement in tiki-taka to control the game. The algorithm also utilizes unsuccessful ball pass and multiple key players to enhance the exploration. MOTTA was tested against popular CEC09 benchmark functions. Findings: The computational experiments indicated that MOTTA had better results in 82% of the cases from the CEC09 benchmark functions. In addition, MOTTA successfully found 83.3% of the Pareto optimal solution in the SALB-E optimization and showed tremendous performance in the spread and distribution indicators, which were associated with the multiple key players in the algorithm. Originality/value: MOTTA exploits the information from all players to move to a new position. The algorithm makes all solution candidates have contributions to the algorithm convergence.