Improved Bat Algorithm for faster convergence in solving optimisation problem
Optimisation is concerned with finding solutions to problems under certain constraints. One of the optimisation approaches is metaheuristic. Metaheuristic algorithms are inspired by nature and utilise intelligent mechanisms. In this study, one of the metaheuristic algorithms known as the Bat Algorit...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English English |
| Published: |
2021
|
| Subjects: | |
| Online Access: | http://eprints.utem.edu.my/id/eprint/26076/1/Improved%20Bat%20Algorithm%20for%20faster%20convergence%20in%20solving%20optimisation%20problem.pdf http://eprints.utem.edu.my/id/eprint/26076/2/Improved%20Bat%20Algorithm%20for%20faster%20convergence%20in%20solving%20optimisation%20problem.pdf http://eprints.utem.edu.my/id/eprint/26076/ https://plh.utem.edu.my/cgi-bin/koha/opac-detail.pl?biblionumber=121246 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Optimisation is concerned with finding solutions to problems under certain constraints. One of the optimisation approaches is metaheuristic. Metaheuristic algorithms are inspired by nature and utilise intelligent mechanisms. In this study, one of the metaheuristic algorithms known as the Bat Algorithm (BA) has been discussed. Previous research has shown that BA is able to provide a good exploration and exploitation in finding solutions. However, this standard BA has the tendency to be trapped in a local minimum when applied to high dimensional search spaces besides experiencing slow convergence rate and low accuracy. The standard BA may be improved by integrating it with additional techniques which can increase its robustness through faster convergence and eventually producing more accurate results. Thus, this study proposed an Improved Bat Algorithm (IBA) by introducing some modifications to the standard BA. The additional techniques included are inertia weight factor, modified new bat position and adaptive boundary size. The IBA is evaluated and tested through a sequence of experiments conducted with ten benchmark functions. For comparison, three established algorithms namely Harmony Search (HS), Particle Swarm Optimisation (PSO) and Genetic Algorithm (GA) are analysed through the same set of experiments and compared with the IBA. The results show that the IBA performs better than Harmony Search (HS), Particle Swarm Optimisation (PSO) and Genetic Algorithm (GA). Despite the high dimensionality of the boundary size, the IBA is still able to produce significant results with the small number of iterations and fast convergence compared to other algorithms. Besides that, IBA was found comparable with existing variants of BA such as the IBA developed from the previous researcher in the year 2013 and the Hybrid Self-Adaptive Bat Algorithm (HSABA) developed in the year 2014. Finally, the developed IBA is found consistent with the exact method which is the simplex method when tested through fairness nurse scheduling problem. Therefore, this confirms the validity of the IBA as an alternative algorithm for solving optimisation problems. |
|---|