Combinatorial optimization with distributed genetic algorithms

With combinatorial optimization we try to find good solutions for many computationaly difficult problems. For this type of problems we often use metaheuristics such as genetic algorithms. We describe complexity classes NP and P and relation between them. We present genetic algorithms with different parallelization models. Main part consists of distributed genetic algorithm implementation using client-server schema and island model. We develop communications protocol and graphical user interface. We analyze several algorithm and distribution parameters and test our implementation using traveling salesman problem collection