割平面算法简介.docVIP

割平面法在纯整数规划中的应用 1、摘要： 割平面法是整数规划问题中常用的一种重要方法。割平面法解整数规划问题的基本思路是：首先根据单纯形法，画出单纯形表格,利用迭代法求出不考虑整数约束条件时的松弛问题的最优解，如果得到的解是整数则即为问题的整数解，运算停止。但是在大多数情况下得到的解不完全是整数，其中会出现非整数的形式，也就是这个松弛问题的最优解中存在某个或者某几个基变量为非整数的形式，这就需要进一步的运算：从最优表格中提取出关于非整数基变量的约束等式，再从这个约束等式出发构造出一个割平面方程添加到原来的约束条件中去，再进行单纯形表格的迭代运算，求出最优解，如果得到的最优解是整数则即为所要求的问题的最优解，运算停止。如果得到的解仍然不完全是整数，就需要继续进行运算，重复以上步骤，一直求解出满足条件的最优解则运算停止。这就是割平面法的整数求解的一般步骤。 Cutting plane method which is used in an integer programming problem is a kind of important method. Cutting plane method solution in integer programming problem is the basic train of thought: first according to the simplex method, draw the simplex form, using iterative algorithm to find the integer constraint conditions do not consider the relaxation of the optimal solution of the problem, given the solution is for the problem that is the integer solutions, stop operations. But in most cases the solution doesnt get completely integer, which could not be the integer form, also is the relaxation of the optimum solution of the existence of a or certain base the variable is not the integer form, this needs further operation: the optimal form from the extract about the base variables the constraint equation integer variables, and again from the constraint equation is constructed on a cutting plane equation added to the original conditions to, and then to simplex form iterative operation, to work out the optimal solution, if we get the optimal solution is required for the integer is the problem of optimal solution, stop operations. If the solution have still not quite integer, it needs to continue operations, repeat the above steps, has been worked out to meet the conditions of the optimal solution is to stop operations. This is cutting plane method of solving the integral general steps. 整数规划的简要介绍 整数规划是指在一类问题中要求其解的全部或者一部分变量为整数的数学规划。在一般的线性规划问题中，所求的最优解可能是整数也可能是分数或小数，但是对于某些特殊的具体问题而言，常常要求最优解必须是整数。例如，所求解是工人的人数，货车的车数，机器的台数，送货的次数等等。为了满足所求解是整数的要求，初看起来好像只要把所求的非整数解用凑整数法、四舍五入法或去尾法化为整数就可以了，但是实际上化为整数后的解不一