Source-linked AI summary
BAS: Beetle Antennae Search Algorithm for Optimization Problems
Xiangyuan Jiang, Shuai Li
TL;DR
Optimization problems motivate meta-heuristic methods, but the paper explores a new beetle-inspired approach for detecting and searching. BAS mimics beetle antennae and random walking, and benchmarks validate its efficacy on two test functions.
Problem
The paper addresses optimization with meta-heuristic search by examining whether beetle searching behavior can inspire an effective algorithm.
Method
BAS mimics beetle antennae, random walking, and directional odour detection through detecting and searching steps with decreasing sensing length and step size.
Results
Across two test functions, BAS approached the global minima; for Goldstein-Price, it obtained fbst = 3.0064 near f∗ = 3.
Takeaways & Limitations
Visualization and numerical results validate BAS efficacy for convergence and local minimum avoidance on the two tested functions.
Takeaways & Limitations
Further parameter selections could be studied to analyze BAS convergence.
Abstract
from arXiv · showhide
Meta-heuristic algorithms have become very popular because of powerful performance on the optimization problem. A new algorithm called beetle antennae search algorithm (BAS) is proposed in the paper inspired by the searching behavior of longhorn beetles. The BAS algorithm imitates the function of antennae and the random walking mechanism of beetles in nature, and then two main steps of detecting and searching are implemented. Finally, the algorithm is benchmarked on 2 well-known test functions, in which the numerical results validate the efficacy of the proposed BAS algorithm.
I. INTRODUCTION
Meta-heuristic algorithms have attracted sustained attention because of their simplicity, flexibility, and ability to avoid local optima. They are important in computer science and practical engineering, with methods inspired by natural behaviors such as grey-wolf leadership and cuckoo brood parasitism.
- Meta-heuristic algorithms have attracted attention over the last decades because of their simplicity, flexibility, and local optimum avoidance.
- These algorithms play an important role in academic computer science and many practical engineering fields.
- Examples include grey wolf optimization, inspired by wolf leadership and prey, and cuckoo-inspired methods based on brood parasitism.
II. BEETLE ANTENNAE SEARCH ALGORITHM DESIGN
The BAS algorithm formulates beetle antennae searching as an optimization process with random exploration, side-by-side sensing, and iterative movement toward higher-fitness regions. Its design uses antennae length and step size to balance local exploitation, escape from local minima, and convergence.
- Behavioral inspiration: BAS models beetle searching through two behaviors: random exploration and detecting odour differences between the two antennae.A beetle explores nearby areas randomly, then turns toward the side detecting higher odour concentration or turns oppositely otherwise.
- Optimization formulation: The algorithm associates beetle position x_t with an objective function f(x), where the maximum fitness identifies the odour source.The beetle position is represented as a vector at time t, while f(x) gives odour concentration and fitness.
- Searching and detecting: BAS generates a random search direction and evaluates right-hand and left-hand antenna positions to imitate beetle sensing.The antennae positions are represented as x_r and x_l, with d defining the sensing length.
- Parameter design: The sensing length d should begin large enough to cover a broad area and then attenuate over time, supporting escape from local minimum points.The parameter controls exploit ability and is intended to balance broad initial search with later refinement.
- Parameter design: The step size δ controls convergence speed and follows a decreasing function of t, with initialization equivalent to the searching area.Both antennae length and step size may alternatively be designated as constants when necessary.
III. BENCHMARK VALIDATION
The BAS algorithm is benchmarked on the Michalewicz and Goldstein–Price functions to verify its efficacy. In both simulations, BAS finds or seeks the functions’ global optima over iterative steps.
- BAS is evaluated on benchmark functions to validate the proposed algorithm’s efficacy.
- For the Michalewicz function, BAS finds the global minimum f∗≈−1.801 at x∗≈(2.20319, 1.57049) in i = 2 dimension.The simulation runs from time step 0 to 100 with d0 = 2 and δ0 = 0.5.
- For the Goldstein–Price function, the global minimum is f(x∗) = 3 at x∗= [0, −1]T, and BAS seeks this optimum through 100 iterative steps.The input domain is usually {x|xi ∈ [−2, 2], i = 1, 2}, and the simulation uses the same parameter configuration as the Michalewicz experiment.
IV. CONCLUSION
The work presents a nature-inspired BAS algorithm that mimics beetle detecting and searching behaviors for optimization. Benchmarking on two typical test functions validates its efficacy in convergence and local minimum avoidance.
- Contribution: BAS is a novel nature-inspired algorithm for solving optimization problems.The method is inspired by beetle behavior.
- Method: The algorithm mimics beetles’ detecting and searching behaviors.Further parameter selection could be studied to improve algorithm convergence.
- Evaluation: Two typical test functions benchmark BAS for convergence and local minimum avoidance, with visualization and numerical results validating its efficacy.Both visualization results and numerical solutions support the proposed algorithm’s effectiveness.