Sami Khuri

• Main Authors: San Jos'e, Sami Khuri, Thomas Back, Jorg Heitkotter
• Other Authors: The Pennsylvania State University CiteSeerX Archives
• Format: Text
• Language: English
• Published: ACM Press 1994
• Subjects: sami
• Online Access: http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.51.2713; http://ls11-www.informatik.uni-dortmund.de/people/baeck/papers/sac94.ps.gz

A genetic algorithm, GENEsYs, is applied to an NPcomplete problem, the 0/1 multiple knapsack problem. The partitioning of the search space resulting from this highly constrained problem may include substantially large infeasible regions. Our implementation allows for the breeding and participation of infeasible strings in the population. Unlike many other GA-based algorithms that are augmented with domain-specific knowledge, GENEsYs uses a simple fitness function that uses a graded penalty term to penalize infeasibly bred strings. We apply our genetic algorithm to problem instances from the literature of well known test problems and report our experimental results. These encouraging results, especially for relatively large test problems, indicate that genetic algorithms can be successfully used as heuristics for finding good solutions for highly constrained NP-complete problems. Genetic Algorithms Direct random search algorithms based on the model of organic evolution received remarka.