Abstract

Metric functions are introduced for various classes of single-machine scheduling problems. It is shown how approximate solutions of NP-hard problems can be found using these functions. The metric value is determined by solving a linear programming problem with constraints being systems of linear inequalities for polynomial or pseudopolynomial solvable instances of the problem under study. In fact, the initial instance is projected onto the subspace of solvable problem instances in the introduced metric.

中文翻译：

---

寻找调度问题近似解的度量方法

摘要

针对各类单机调度问题引入了度量函数。展示了如何使用这些函数找到NP难题的近似解。度量值通过求解线性规划问题来确定，其中约束是所研究问题的多项式或伪多项式可解实例的线性不等式系统。事实上，在引入的度量中，初始实例被投影到可解决问题实例的子空间上。