Enhanced Ant Colony Optimization with Dynamic Mutation and Ad Hoc Initialization for Improving the Design of TSK-Type...

Area of Science:

• Computational Intelligence
• Fuzzy Systems Engineering
• Optimization Algorithms

Background:

• Takagi-Sugeno-Kang (TSK) fuzzy systems are widely used in control and prediction.
• Existing population-based algorithms often use generic initialization, limiting design accuracy.
• Improving the accuracy and convergence of TSK fuzzy system design is a key challenge.

Purpose of the Study:

• To propose an enhanced ant colony optimization algorithm, ACODM-I, for improving TSK fuzzy system design.
• To introduce an ad hoc, application-specific initialization strategy for initial ant solutions.
• To incorporate dynamic mutation into continuous ACO (ACO_R) to balance exploration and convergence.

Main Methods:

• Developed ACODM-I with ad hoc initialization and dynamic mutation.
• Integrated dynamic mutation into the existing ACO_R algorithm.
• Validated the algorithm through simulations on TSK fuzzy systems for plant tracking and time series prediction.

Main Results:

• ACODM-I demonstrated superior performance compared to ACO_R and other advanced algorithms.
• The proposed ad hoc initialization significantly improved fuzzy system design accuracy.
• Dynamic mutation enhanced the exploration ability and convergence rate of the optimization process.

Conclusions:

• ACODM-I offers a superior approach for designing accurate TSK fuzzy systems.
• The combination of ad hoc initialization and dynamic mutation is effective in improving optimization.
• The algorithm shows promise for applications in dynamic control and chaotic time series prediction.