间歇化工过程的最优设计问题是一类复杂且难以求解的组合优化问题。通过把这类问题分解为只包含离散变量的主导问题和只含连续变量的子问题，把遗传算法和线性规划法结合起来对其进行求解。并在算法中引入了一类新的算子，显著地提高了收敛概率。算例表明，该方法可以避免直接求解过程的复杂性和困难，并且具有很好的全局收敛性。 The optimal design of batch chemical processes is a complex and difficult to solve combinatorial optimization problem. By decomposing this kind of problem into a dominant problem containing only discrete variables and a sub-problem containing only continuous variables, genetic algorithms and linear programming methods are combined to solve them. In the algorithm, a new kind of operator is introduced, which significantly improves the convergence probability. The example shows that this method can avoid the complexity and difficulty of the direct solution process and has good global convergence.