优化设计的设计变量可以为连续型或离散型(整型)。连续变量的优化理论己较成熟,离散变量的优化则正在研讨中。下面给出的是一种先将离散型变量当连续型处理,用连续变量的优化方法得出优化结果后,再将结果数据进行圆整处理,从而得到离散型交量优化解的离散变量圆整方法。一、函数的灵敏度它是指设计变量发生微小变化时,目标函数的变化程度。在函数的一阶偏导数存在时,可定义为函数:f=f(X_1,X_2,…X_n)对变量 X_1在点D~*=(X_1,X_2,…X_n)~*处的灵敏度为在函数的偏导数解析表达式复杂,难以计算时,可以用差商形式定义 Design variables for optimization can be continuous or discrete (integer). Optimization theory of continuous variables has been more mature, optimization of discrete variables is being studied. The following is a first discrete variables as a continuous type, with the continuous variable optimization method to obtain the optimal results, and then the results of the data are rounded to obtain the discrete variable cross-flow optimization solution of the discrete variables The whole method. First, the sensitivity of the function It refers to the small changes in the design variables, the objective function of the degree of change. When the first-order partial derivative of a function exists, it can be defined as a function: f = f (X_1, X_2, ... X_n) The sensitivity of the variable X_1 at point D * = (X_1, X_2, ... X_n) Partial Derivative Function Analysis Expressions complex, difficult to calculate, you can use the form of the definition of the difference