An Improved Algorithm for Solving Scheduling Problems by Combining Generative Adversarial Network with Evolutionary Algorithms

With1 the continuous application of evolutionary algorithms in various combinatorial optimization problems, the traditional evolutionary algorithms are prone to premature convergence and fall into local optimization solutions as the complexity of the problems increases. To solve this problem, this paper proposes a hybrid algorithm combining the Generative adversarial nets (GAN) and Genetic Algorithm (GA). The algorithm is based on Genetic Algorithm and introducted the GAN sample as another sample to the generated model. The algorithm expected more abundant sample information through GAN mining, got the advantage of sample training GAN through the GA. It makes GAN learn from the edge of sample information, which can generate more advantages of samples. The generated sample is injected into the evolution of the next generation, increasing the diversity of samples and increasing the opportunity to find the optimal solution. In this paper, the hybrid algorithm is used to solve the Permutation Flow Shop Problem to verify the algorithm's solution ability. Experimental results show that the hybrid algorithm can avoid premature local optimal solution compared with the traditional evolutionary algorithm.