Distributed genetic algorithms with a new sharing approach in multiobjective optimization problems

In this paper, a new distributed genetic algorithm for multiobjective optimization problems is proposed. In this approach, the island model is used with a distributed genetic algorithm and an operation of sharing for Pareto-optimum solutions is performed with the total population. In multiobjective optimization problems, the Pareto-optimum solutions should be derived for designers. Because the Pareto-optimum solutions are the set of optimum solutions that are in the relationship of trade-off, not only the accuracy but also the diversity of the solutions should be high. The effect of the distributed populations leads to the high accuracy and the sharing effect leads to the high diversity of solutions. These effects are examined and discussed through some numerical examples that have more than three objective functions.