An Improved Whale Optimization Algorithm Based on Different Searching Paths and Perceptual Disturbance

Wei-zhen Sun, Jie-Sheng Wang, Xian WeiSymmetry 2018 · OpenReview

TL;DR

This paper introduces an improved Whale Optimization Algorithm (WOA) called CP-PDWOA, which enhances local search capabilities by incorporating a "perceptual disturbance" mechanism and exploring various spiral search paths. The proposed algorithm demonstrates significantly better performance in terms of convergence speed and accuracy compared to the standard WOA, with the Archimedes spiral proving to be the most effective search path.

What Problem Does This Paper Address?

Optimization problems, which involve finding the best solution from a set of possibilities, are ubiquitous in science and engineering. Swarm intelligence algorithms, inspired by natural collective behaviors like those of ants or whales, are popular for tackling complex optimization tasks. The Whale Optimization Algorithm (WOA), inspired by humpback whale hunting, is known for its strong global search capabilities. However, the original WOA has a deficiency in its local search mechanism, which can lead to premature convergence or getting stuck in local optima. Additionally, its reliance on a logarithmic spiral search path can result in incomplete exploration of the search space. This paper aims to address these limitations by improving WOA's local search and exploring more effective search paths.

Methods

The researchers propose two main improvements to the standard WOA:

  • Perceptual Disturbance Mechanism: This mechanism introduces a controlled "perturbation" around the best-found solutions. This allows the search agents to explore the immediate vicinity of promising areas more thoroughly, helping to avoid local optima and refine the search for the global optimum. The disturbance factor is carefully designed to ensure accuracy and prevent excessive randomness.
  • Complex Search Paths: The original WOA uses a logarithmic spiral for its search path. The authors argue that this path's variable pitch can lead to missed areas in the search space. They explore seven alternative spiral curves, including Archimedes spirals, Rose spirals, Epitrochoids, and Lituus spirals, to find a more effective search trajectory. The equal-pitch Archimedes spiral is specifically highlighted for its potential to improve exploration.

The improved algorithm, termed CP-PDWOA (Complex Path-Perceptual Disturbance WOA), combines these two enhancements. The algorithm's procedure involves initialization, realization of the improved search path (either shrinking encircling or spiral updating), location updates, and the application of the perceptual disturbance to the best agent.

Datasets

The paper evaluates the performance of the proposed CP-PDWOA algorithm using a standard set of 23 benchmark test functions. These functions are categorized into unimodal, multimodal, and fixed-dimension multimodal types, designed to test various aspects of optimization algorithms, including exploitation capability (unimodal) and exploration capability (multimodal). The authors do not introduce any new datasets.

Key Insights and Ideas

  • Perceptual disturbance is crucial for local search: The introduction of a controlled disturbance mechanism significantly improves the algorithm's ability to refine solutions and escape local optima, leading to better convergence and accuracy.
  • Search path matters: The choice of search path has a substantial impact on an algorithm's performance. The variable pitch of the logarithmic spiral can be a bottleneck, whereas a more consistent or structured path can lead to more thorough exploration.
  • Archimedes spiral shines: Among the tested spiral curves, the equal-pitch Archimedes spiral demonstrated superior performance, particularly in improving convergence speed and accuracy for several complex functions.
  • Avoiding local minima is key: The paper emphasizes that effective avoidance of local minima is a critical factor for successful optimization, and their proposed disturbance mechanism directly addresses this challenge.

Key Results

The simulation results, conducted using MATLAB, show that the proposed CP-PDWOA algorithm significantly outperforms the standard WOA.

  • Improved Convergence: The convergence curves clearly indicate that the CP-PDWOA converges faster and to better (lower) objective function values than the original WOA, especially when the perceptual disturbance is applied.
  • Search Path Effectiveness: The Archimedes spiral curve, when used as the search path in the improved WOA, consistently yielded the best results across various benchmark functions, outperforming other spiral curves.
  • Specific Function Improvements: For functions like F5, F6, F10, and F13, the Archimedes spiral path led to a dramatic improvement in convergence speed and accuracy, with the algorithm converging much closer to the optimal value (e.g., near zero for F5) compared to the original WOA.
  • Handling Complex Functions: For complex, multi-peak functions (e.g., F8, F9, F14), the added perturbation mechanism was effective in avoiding local minima and improving the overall search performance.

Limitations and Future Work

  • Parameter Tuning: The paper mentions that randomly defining different search paths for different agents in the same iteration can lead to better performance but also introduces complexity in parameter adjustment.
  • Specific Function Anomaly (F8): The authors note a special case with function F8 where their algorithm, along with others, did not achieve the minimum value reported in a previous study. They acknowledge that the specific reason for this discrepancy was not found.
  • Future Work: While not explicitly detailed as future work, the exploration of various spiral curves suggests an ongoing interest in optimizing the search path. The successful integration of perceptual disturbance also opens avenues for exploring different types and applications of such mechanisms. The availability of the MATLAB code on GitHub suggests a desire for community engagement and further development.

Summary updated

Abstract

Whale optimization algorithm (WOA) is a swarm intelligence optimization algorithm inspired by humpback whale hunting behavior. WOA has many similarities with other swarm intelligence algorithms (PSO, GWO, etc.). WOA’s unique search mechanism enables it to have a strong global search capability while taking into account the strong global search capabilities. In this work, considering the the deficiency of WOA in local search mechanism, combined with the optimization methods of other group intelligent algorithms, perceptual perturbation mechanism is introduced, which makes the agent perform more detailed searches near the local extreme point. At the same time, since the WOA uses a logarithmic spiral curve, the agent cannot fully search all the spaces within its search range, even though the introduction of the perturbation mechanism may still lead to the algorithm falling into a local optimum. Therefore, the equal pitch Archimedes spiral curve is chosen to replace the classic logarithmic spiral curve. In order to fully verify the effect of the search path on the performance of the algorithm, several other spiral curves have been chosen for experimental comparison. By utilizing the 23 benchmark test functions, the simulation results show that WOA (PDWOA) with perceptual perturbation significantly outperforms the standard WOA. Then, based on the PDWOA, the effect of the search path on the performance of the algorithm has been verified. The simulation results show that the equal pitch of the Archimedean spiral curve is best.
BibTeX
@inproceedings{sun2018improved,
  author = {Wei-zhen Sun and Jie-Sheng Wang and Xian Wei},
  title = {{An Improved Whale Optimization Algorithm Based on Different Searching Paths and Perceptual Disturbance}},
  booktitle = {Symmetry},
  year = {2018},
  url = {https://openreview.net/forum?id=Vn7Dpa0v8V},
}
Figure