A Complex Chained P System Based on Evolutionary Mechanism for Image Segmentation.

Area of Science:

• Computational intelligence
• Membrane computing
• Bio-inspired computing

Background:

• Clustering problems require efficient algorithms for data partitioning.
• Existing Particle Swarm Optimization (PSO) methods have limitations in global search ability and premature convergence.
• Membrane systems offer a framework for parallel and distributed computation.

Purpose of the Study:

• To design and implement a novel Complex Chained P system (CCP) for enhanced clustering.
• To improve global search capabilities and avoid premature convergence in clustering algorithms.
• To validate the effectiveness of CCP in solving real-world clustering and image segmentation tasks.

Main Methods:

• Development of a Complex Chained P system (CCP) integrating evolutionary mechanisms.
• Utilization of two types of evolution rules: Particle Swarm Optimization (PSO) and Differential Evolution (DE).
• Incorporation of communication rules to accelerate convergence and prevent premature solutions.
• Implementation of a distributed parallel computing model for enhanced processing.

Main Results:

• The proposed CCP demonstrated superior performance compared to four existing PSO clustering approaches.
• Validation was conducted on eight real-life datasets, showing significant improvements.
• Computational results confirmed CCP's effectiveness in image segmentation problems.

Conclusions:

• The Complex Chained P system (CCP) offers a robust and effective solution for clustering problems.
• The integration of PSO and DE within the CCP framework enhances global search and convergence.
• CCP shows significant potential for applications in image segmentation and other complex data analysis tasks.