An Artificial Immune System to Handle Constraints

Nareli Cruz Cortés developed a hybrid between a genetic algorithm and an artificial immune system to solve constrained (single-objective) optimization problems. This approach does not add any significant computational cost the the simple genetic algorithm, and considerably improves its performance when solving constrained optimization problems. The main idea is to evolve another genetic algorithm (within the main GA used to solve a problem) using as fitness a Hamming distance between the "antigens" (a reference solution that we want to match) and the "antibodies" (the solutions that we want to improve).


The source code of this approach is available here.


For further information about this approach, please refer to:

1. Carlos A. Coello Coello and Nareli Cruz Cortés, Hybridizing a Genetic Algorithm with an Artificial Immune System for Global Optimization, Engineering Optimization, Vol. 36, No. 5, pp. 607--634, October 2004.
   
   
   
2. Coello Coello, Carlos A. and Cruz Cortés, Nareli, A Parallel Implementation of an Artificial Immune System to Handle Constraints in Genetic Algorithms: Preliminary Results, Congress on Evolutionary Computation (CEC'2002), IEEE Service Center, Piscataway, New Jersey, Volume 1, pp. 819--824, May 2002.
   
   
   
3. Coello Coello, Carlos A. and Cruz Cortés, Nareli, Use of Emulations of the Immune System to Handle Constraints in Evolutionary Algorithms, in Cihan H. Dagli, Anna L. Buczak, Joydeep Ghosh, Mark J. Embrechts, Okan Erson and Stephen Kercel (eds.), Intelligent Engineering Systems through Artificial Neural Networks (ANNIE'2001), ASME Press, Vol. 11, pp. 141--146, St. Louis Missouri, July 2001 (best paper award).