Multi- Populations Genetic Algorithms for Vehicle Routing Problems
Vehicle Routing Problems, how namely use the limited transportation resources to complete the ration the transportation duty, and causes the transportation cost lowest question. Vehicle Routing Problems as a result of its huge economic efficiency, obtained the development during more than 40 years in the past which progresses by leaps and bounds. This article in by has in the foundation which the method studies, is easy in view of the standard genetic algorithms in the solution Vehicle Routing Problems to appear precociously, is easy to fall into the partial optimal solution shortcoming, makes the improvement to the traditional genetic algorithms, proposed the multi- populations genetic algorithms. In the solution process the initialization two populations, separately will select the different overlapping variation probability, after each time will iterate the sufficiency high individual carries on the first populations in sufficiency low individual with the second population in the exchange, and will preserve each center group the optimal solution to the outstanding person population, by will solve the tradition genetic algorithms to be easy to appear precociously, will be easy to fall into the partial optimal solution question. The experimental result indicated that, after improvement genetic algorithms compared to general algorithm convergence rate quicker, the solution quality is finer.
Key word: Genetic Algorithms, Vehicle Routing Problems, Good Population and Bad Population, Elite Population