An Adaptive Genetic Algorithm for Digital Facility Layout Optimization Under the Multi-Species and Variable-Batch Production Mode

Abstract: For the NP-hard characteristic of facility layout problem (FLP) and its importance to industry, many optimal and heuristic algorithms have been designed to solve the problem. But when optimizing the production, with fixed probabilities, traditional GA has its flaws with slow convergence speed and the less-than ideal accuracy of the optimal solution. According to the characteristics of multi-species and variable-batch production mode, this paper analyzed those weak points and proposed an improved adaptive genetic algorithm with the objective of minimizing the material handling cost. This genetic algorithm has adaptive probabilities based on the fitness. It can keep the diversity and excellence of the genes. The proposed model of GA also has been tested and verified by simulation with MTALAB. According to the results, it shows that the proposed approach improved the convergence speed and the accuracy of the optimal solution and can help the factory make right decision.