Improved Genetic Algorithm And Its Application In Mould Optimization Design

Genetic Algorithm (GA) is a high parallel, random and adaptive global optimizationalgorithm, which is inspired from the mechanism of nature selection and evolution. Itsadvantage is the effectivity in resolving complicated and non-linear problems, which aredifficult for traditional searching methods. For many characteristics, i.e., few limitations onthe presumption of the solution space and better abilities of robustness, it has been widely andsuccessfully applied in many areas such as function optimizations, machine learning andself-adaptive control systems, etc.Mould optimization design is an international front problem and interesting areacurrently. In the optimization process, the object and constraint functions are implicitfunctions, which make sensitivity analysis and functions calculation become difficultextremely, so it is important to study high performance optimization method of mould design.The main work of this paper is researching the methods to improve GA and applyingthem in the mould optimization design.Firstly, the paper summarizes the histories and developments of mould optimizationdesign and GA, and introduces the applications, basic theory and technical implement of GA.Secondly, a Multi-population Genetic Algorithm (KSPGA) based on Species equationand Kriging operator is presented in this paper. The parameters of species equation areconsidered as design variables, the stable solution of equation is regarded as modifiedarithmetic crossover operator to participate in genetic operation, which can improveindividual varieties. The Kriging operator which can simulate best solution through all theindividuals information is brought in to enhance the ability of searching optimal solution.Immigration operator can increase communion of populations and promote convergence.Numerical examples show the higher efficiency and better applicability of the KSPGA.Finally, the KSPGA, combined with Z-MOLD simulation programs, is used to search theoptimal gate locations. The results indicate the KSPGA is effective in obtaining the optimalgate location for single-gate mold. It plays a positive role in the improvements of mouldoptimization technology and manufacture level.