The partial representation extension problem generalizes the recognition problem for classes of graphs defined in terms of vertex representations. We exhibit circular-arc graphs as the first example of a graph class where the recognition is polynomially solvable while the representation extension problem is NP-complete. In this setting, several arcs are predrawn and we ask whether this partial representation can be completed. We complement this hardness argument with tractability results of the representation extension problem on various subclasses of circular-arc graphs, most notably on all variants of Helly circular-arc graphs. In particular, we give linear-time algorithms for extending normal proper Helly and proper Helly representations. For normal Helly circular-arc representations we give an $O(n^3)$-time algorithm. Surprisingly, for Helly representations, the complexity hinges on the seemingly irrelevant detail of whether the predrawn arcs have distinct or non-distinct endpoints: In the former case the previous algorithm can be extended, whereas the latter case turns out to be NP-complete. We also prove that representation extension problem of unit circular-arc graphs is NP-complete.
翻译:部分代表扩展问题 部分代表扩展问题 概括了以顶点表示方式定义的图表类别的识别问题 。 我们展示循环弧图作为图表类别的第一个示例, 其识别方式是多角度的可溶性, 而代表扩展问题是完整的 。 在这种背景下, 有几个弧是预草的, 我们问这个部分代表方式能否完成 。 我们用循环弧图各小类中代表扩展问题的可移植性结果来补充这种硬性论点, 最突出的是 Helly 圆弧图的所有变量 。 特别是, 我们给出线性时间算法, 用于扩展正常的正常的 Helly 和适当的 Helly 表达方式 。 对于普通的 Hellly 循环弧- 表达方式, 我们给出了一个$( n) 3 美元- 时间算法 。 令人惊讶的是, 复杂程度取决于预先弧弧是否具有独特或非模糊的终点的表面细节 : 在前一种情况下, 以前的算法可以扩展, 而后一种情况则转换为 NP- 完整 。 我们还证明圆形图的扩展 。