Gravitational search – bat algorithm for solving single and bi-objective of non-linear functions

In this thesis, in order to solve single objective optimization problem and bi-objective objective optimization problem in non-linear functions, two methods are created during the course of the present work. Firstly, a new strategy based on a combined method (i.e. single-objective Gravitational Search (GSA) with Bat Algorithm (BAT) (SOGS-BAT)) algorithm is proposed in which relies on the closed interval between 0 and 1 to avoid falling into local search. The lack of local optimum mechanism decreases the intensification of the search space, whereas diversity remains high. Secondly, two meta-heuristics, namely, Bi-Objective Gravitational Search Algorithm (BOGSA) and Bi-Objective Bat Algorithm (BOBAT), were combined to form a (BOGS-BAT) algorithm. Later, this algorithm was used to solve bi-objective Production Planning (PP) and Scheduling Problem (Sch.P). The BOGS-BAT algorithm is based on three techniques. The first technique is to move or switch solution from single function to functions that contain more than one objective functions. The use of the BOGSA algorithm aims to create a new equation for the calculation of the masses of population individuals, as found in the theoretical work in the Strength Pareto Evolutionary Algorithm two (SPEAII) algorithm. The second technique is to solve bi-objective functions by using the BOBAT algorithm. The third technique is an integration of BOGSA with BOBAT to produce a BOGSBAT algorithm. The gravitational search with BAT algorithm is used to balance exploitation and exploration, thereby resulting in efficient and effective (speed and accuracy) solution for the production planning model. Finally, to verify the efficiency of the SOGS-BAT and BOGS-BAT and to demonstrate the effectiveness and robustness of the proposed algorithms, the numerical experiments based on benchmark test functions were performed. In addition, the simulation random data for were used to solve single and bi-objective optimization PP and Sch.P to improve the validation and verify the performance of the proposed algorithms. The results reveal that the proposed algorithms are promising and efficient.

Text FS 2019 27 - ir.pdf Download (1MB)

Actions (login required)

View Item