BAS-ADAM: an ADAM based approach to improve the performance of beetle antennae search optimizer

Abstract

In this paper, we propose enhancements to Beetle Antennae Search (BAS) algorithm, called BAS-ADAM, to smoothen the convergence behavior and avoid trapping in local-minima for a highly non-convex objective function. We achieve this by adaptively adjusting the step-size in each iteration using the Adaptive Moment Estimation (ADAM) update rule. The proposed algorithm also increases the convergence rate in a narrow valley. A key feature of the ADAM update rule is the ability to adjust the step-size for each dimension separately instead of using the same step-size. Since ADAM is traditionally used with gradient-based optimization algorithms, therefore we first propose a gradient estimation model without the need to differentiate the objective function. Resultantly, it demonstrates excellent performance and fast convergence rate in searching for the optimum of non-convex functions. The efficiency of the proposed algorithm was tested on three different benchmark problems, including the training of a high-dimensional neural network. The performance is compared with Particle Swarm Optimizer (PSO) and the original BAS algorithm.

Randomized time‐varying knapsack problems via binary beetle antennae search algorithm: Emphasis on applications in portfolio insurance

“…Beetle search behavior can be designed to solve the problem of optimizing an objective function. This approach allows the implementation of new optimization algorithms (see other studies [33][34][35] ).…”

“…The advantage of the online solution in a time interval is that it includes much less noise than if we just take all the solutions of every problem in that interval and merge them. BAS is a meta-heuristic algorithm for intelligent optimization processes commonly used in many scientific fields (see, for example, other studies [17][18][19] ), and it is directly applicable only to unconstrained optimization. Here we have modified the BAS algorithm into a Binary BAS (BBAS) algorithm that has been translated into a MATLAB setup to handle ILP problems.…”

Mobile Robot Path Planning Based on Time Taboo Ant Colony Optimization in Dynamic Environment

“…However, all kinds of algorithms are more or less limited to algorithm deficiencies. Compared with the traditional gradient descent algorithm, the metaheuristic algorithm performs better in convergence speed, and global optimization ability and hence is widely used in trajectory planning (Khan et al, 2020b ), prediction, resource scheduling, and other fields. Ant colony algorithm with its good robustness, positive feedback, and parallel computing ability has been widely used in robot path planning and achieved good results.…”

Validation of Large-Scale Classification Problem in Dendritic Neuron Model Using Particle Antagonism Mechanism

“…For example, the gravitational search algorithm is a basic learning algorithm to simulate physical phenomena [25][26][27][28][29], biogeography-based optimization (BBO) [30] is generally used for simulating ecological concepts since the accuracy and stability are the most outstanding among the models using representative metaheuristics [31], and some basic learning algorithms can simulate the moving sample population of organisms such as particle swarm optimization (PSO) [32,33] and ant colony optimization. Moreover, as a variant of PSO, the competitive swarm optimizer (CSO) [34,35] is a simplified metaheuristics set that is suitable for both multi-point and local exploration. Compared to the systems hat only conduct multi-point exploration or local exploration, the trap of the local optimal solution and convergence rate can be balanced using the CSO.…”