Research on Solving Nonlinear Integer Programming Based on Multi-agent Genetic Algorithm

To solve nonlinear integer programming which is widely used in engineering fields, a multi-agent genetic algorithm (MAGA) is put forward. In MAGA, an agent represents not only a candidate solution to the optimization problem but also an individual to GA. One hand agents live in society of agents, which can act as agents to achieve their goals; the other hand each agent as an individual of genetic algorithm can carry on the genetic operators to achieve their goals. After the numerical experiment, it is proved to be feasible and effective, especially in solving large-scale problems, it shows much better performance.