Artificial dragonfly algorithm in the Hopfield neural network for optimal Exact Boolean k satisfiability...

Area of Science:

• Computational Intelligence
• Artificial Intelligence
• Optimization Algorithms

Background:

• Exact Boolean k-Satisfiability (EBkSAT) problems are computationally challenging.
• Existing methods for EBkSAT representation often face limitations in training speed and accuracy.
• Optimizing logical rule representation is crucial for various computational tasks.

Purpose of the Study:

• To introduce a novel hybrid computational approach integrating the Artificial Dragonfly Algorithm (ADA) with the Hopfield Neural Network (HNN).
• To investigate the effectiveness of ADA in accelerating HNN training for optimized EBkSAT logic representation.
• To evaluate the performance and robustness of the proposed ADA-HNN-EBkSAT model.

Main Methods:

• Development of a hybrid ADA-HNN computational model.
• Construction of a specific EBkSAT problem instance with simulated datasets for evaluation.
• Performance assessment using metrics such as global minimum ratio (GmR), RMSE, MAPE, and computational time (CT).

Main Results:

• The proposed ADA-HNN-EBkSAT model demonstrated superior accuracy compared to existing methods.
• The hybrid model significantly reduced the computational time required for EBkSAT representation.
• Comparative analysis confirmed the effectiveness and robustness of the ADA algorithm with HNN.

Conclusions:

• The ADA-HNN-EBkSAT hybrid model offers an effective and efficient solution for optimal EBkSAT logic representation.
• ADA shows strong compatibility with HNN, enhancing training speed and solution quality.
• This approach has significant implications for solving complex optimization problems in computer science, engineering, and business.