Using genetic algorithms to solve combinatorial optimization problems
Genetic algorithms are stochastic search techniques based on the mechanics of natural selection and natural genetics. Genetic algorithms differ from traditional analytical methods by using genetic operators and historic cumulative information to prune the search space and generate plausible solution...
| Main Author: | |
|---|---|
| Format: | Others |
| Published: |
FIU Digital Commons
1991
|
| Subjects: | |
| Online Access: | http://digitalcommons.fiu.edu/etd/2684 http://digitalcommons.fiu.edu/cgi/viewcontent.cgi?article=3893&context=etd |
| Summary: | Genetic algorithms are stochastic search techniques based on the mechanics of natural selection and natural genetics. Genetic algorithms differ from traditional analytical methods by using genetic operators and historic cumulative information to prune the search space and generate plausible solutions. Recent research has shown that genetic algorithms have a large range and growing number of applications.
The research presented in this thesis is that of using genetic algorithms to solve some typical combinatorial optimization problems, namely the Clique, Vertex Cover and Max Cut problems. All of these are NP-Complete problems. The empirical results show that genetic algorithms can provide efficient search heuristics for solving these combinatorial optimization problems.
Genetic algorithms are inherently parallel. The Connection Machine system makes parallel implementation of these inherently parallel algorithms possible. Both sequential genetic algorithms and parallel genetic algorithms for Clique, Vertex Cover and Max Cut problems have been developed and implemented on the SUN4 and the Connection Machine systems respectively. |
|---|