Constrained optimization via artificial immune system.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Bio-inspired Computing

Background:

• Constrained optimization problems are prevalent in various scientific and engineering fields.
• Existing evolutionary computation methods face challenges in efficiently handling complex constraints.
• Biological immune systems offer a robust framework for adaptation and problem-solving.

Purpose of the Study:

• To propose a novel artificial immune system (AIS) for solving constrained optimization problems.
• To draw an analogy between biological immune responses and constrained optimization.
• To develop an effective optimization strategy by integrating feasible and infeasible solution exploration.

Main Methods:

• Classifying population into feasible and infeasible groups based on constraint violations.
• Employing immune mechanisms like clonal selection, recombination, and hypermutation for feasible solutions.
• Utilizing location updates and direction information for infeasible solutions to explore the feasibility boundary.
• Validating the proposed AIS using recent benchmark functions and comparing with state-of-the-art methods.

Main Results:

• The proposed artificial immune system demonstrates effective handling of constrained optimization problems.
• The approach shows competitive and promising performance compared to various evolutionary computation paradigms.
• The integration of feasible and infeasible group strategies enhances exploration and exploitation balance.

Conclusions:

• The artificial immune system provides a viable and effective approach for tackling constrained optimization.
• The bio-inspired strategy offers a promising alternative to existing optimization techniques.
• Further research can explore the application of this AIS to more complex real-world optimization challenges.