Constrained minimization utilizing GA based pattern recognition of immune system

Abstract

The immune system has pattern recognition capabilities based on reinforced learning, memory and affinity maturation interacting between antigens and antibodies. The paper deals with the adaptation of artificial immune system into genetic algorithm based design optimization. The present study utilizes the pattern recognition from the immune system and the evolution from genetic algorithm. The basic idea is derived from the fact that designs should be distinguished whether they are usable/feasible or infeasible and should be improved towards the optimal solution. For the expression of design solutions, binary coded strings are used to represent antigens and antibodies in artificial immune system and chromosomes in genetic algorithm. The paper discusses the procedure of constrained optimization that does not rely on any detailed mathematical formulation for constraint handling. A number of mathematical function minimization problems are examined for verification, and practical engineering optimization problems including inequality constraints are explored to support the proposed strategy.