摘要
考虑制造工艺要求,将所有设计变量均视为离散变量(包括一般离散变量和伪离散变量),并就这两种情况下状态产生函数的设计原理进行深入研究,解决了将模拟退火算法用于离散变量函数优化的关键技术问题,介绍了一种基于模拟退火算法的离散变量函数优化的新方法。建立了以壳体质量最轻为设计目标、以内径(公称直径)和名义壁厚为设计变量的优化设计的数学模型。该模型具有较多的局部极值点,用以数学规划理论为基础的经典约束优化方法求解效果较差,用基于模拟退火算法的离散变量优化设计方法可以取得良好的效果。
Taking request in processing technique into account, all design variables are regarded as discrete variables, comprising two kinds of general discrete variables and dummy discrete variables, design philosophy of neighborhood function are expounded profoundly in different cases, and a bottle - neck problem to apply simulated annealing algorithm to discrete variables function optimization is solved, a new method of discrete variables function optimization based on simulated annealing algorithm is presented. The mathematical model of optimum design for the pressure vessel is derived. This model takes the minimum mass of shell as objective function, and takes internal diameter--nominal diameter and nominal thickness as the design variables. It has a lot of local minimums, therefore to discover global minimum is very hard. As an example of application, the method has been successfully used to handle the optimum design of the pressure vessel.
出处
《压力容器》
北大核心
2007年第2期15-18,26,共5页
Pressure Vessel Technology
基金
江苏省高校自然科学研究计划项目(04KJD460025)
关键词
压力容器
模拟退火算法
离散变量
优化设计
pressure vessel
simulated annealing algorithm
discrete variables
optimum design
