A flexible annealing chaotic neural network to maximum clique problem

Area of Science:

• Computational Neuroscience
• Artificial Intelligence
• Optimization Algorithms

Background:

• Optimization problems are computationally challenging.
• Existing neural network models may lack flexibility or convergence speed.
• Chaotic dynamics offer potential for complex problem-solving.

Purpose of the Study:

• To introduce a novel flexible annealing chaotic neural network.
• To enhance control over chaotic dynamics for optimization.
• To improve convergence rate and solution quality in optimization tasks.

Main Methods:

• Analysis and comparison of various annealing strategies.
• Development of a flexible annealing chaotic neural network architecture.
• Testing on the maximum clique problem using diverse graph instances (DIMACS, p-random, k-random).

Main Results:

• The flexible annealing chaotic neural network exhibits adjustable chaotic dynamics.
• The network demonstrates quick convergence to stable states.
• Satisfactory solutions for the maximum clique problem were achieved rapidly.
• Superior executive efficiency and near-optimal solutions compared to other chaotic neural networks.

Conclusions:

• The proposed network offers flexible control and fast convergence for optimization.
• It provides a robust and efficient method for solving complex problems like the maximum clique problem.
• The network outperforms existing chaotic neural networks in efficiency and solution quality.