Optimisation of energy efficient Assembly Sequence Planning using Moth-Flame Optimisation alghorithm

• Main Author: Muhammad Arif, Abdullah
• Format: Thesis
• Language: English
• Published: 2019
• Subjects: TJ Mechanical engineering and machinery
• Online Access: http://umpir.ump.edu.my/id/eprint/30029/1/Optimisation%20of%20energy%20efficient%20assembly%20sequence%20planning.pdf; http://umpir.ump.edu.my/id/eprint/30029/
• Institution: Universiti Malaysia Pahang Al-Sultan Abdullah

In assembly optimisation, Assembly Sequence Planning (ASP) refers to the activity of finding the best possible assembly sequences developed on the foundation of assembly modelling. This problem is a non-deterministic polynomial-time (NP) hard problem, as currently cannot be solved by using a specific approach. By considering the complexity to optimise the ASP problem, the traditional approach that evaluates every single possible solution for ASP is inconvenient to be used due to time constraint, high energy consumed and computational cost. In ASP, the research on problem optimisation is important and needs an effective computational approach to determine the best assembly sequence. Research on ASP has been given a lot of attention, especially with the profit and humanrelated objectives. However, based on the literature survey, less attention was given to tackle the sustainable issue in assembly such as carbon emission and energy utilisation. On the other hand, the recent ASP research tends to explore the potential of a relatively new algorithm to optimise ASP. Therefore, the aim of this research is to establish a methodology and implement the relatively new algorithm to optimise the Energy Efficient Assembly Sequence Planning (EE-ASP) problem. In the proposed model, the idle energy utilisation was optimised together with the assembly direction and tool changes. For optimisation purpose, this research proposed a relatively new algorithm called the Moth-Flame Optimisation (MFO). A computational experiment was performed by using the six test problems from the literature. Furthermore, a case study was conducted to validate the proposed EE-ASP model and the performance of the optimisation algorithms. The MFO performance was compared with three frequently used meta-heuristics algorithms in ASP, namely Ant Colony Optimisation (ACO), Genetic Algorithm (GA) and Particle Swarm Optimisation (PSO). Based on the computational experiment, MFO achieved the best results in terms of minimum fitness, maximum fitness, average fitness, and standard deviation in each test problem. Meanwhile, the case study results also indicated that MFO obtained the best minimum fitness, maximum fitness and average fitness. All of the results were then analysed by using the Analysis of Variance (ANOVA) method. The P-value was found to be lower than the significant level α (P ≤ 0.10) in all test and case study problems. Therefore, it could be interpreted that there were statistically significant differences among the group means. The outcomes of ANOVA for the test and case study problems were further analysed with the post-hoc Fisher’s Least Significant Difference (LSD) technique. The LSD result indicated that the MFO had a significant difference in 67 % of the cases as compared to the comparison algorithms. The result from the case study confirmed that the proposed EE-ASP model and MFO algorithm were applicable for the actual assembly data. The proposed MFO layout was able to reduce the idle energy utilisation up to 11.7 %, while the direction change and tool change reduced to 26.67 % and 13.64 % respectively. The findings from this research concluded that the idle energy utilisation model for ASP can be used as a guideline to design a station for sustainable assembly process. Besides that, the MFO had a great potential to be further explored to optimise the combinatorial problem.