我认为在大多数情况下,查看图G的拉普拉斯算子会更有效率,它与邻接矩阵密切相关。在这里,您可以使用它来将第二个特征值与图的“局部与全局”属性相关联。
为简单起见,我们假设G为d正则。则G的规范化Laplacian 为L=I−1dA,其中I是n×n恒等式,A是邻接矩阵。关于拉普拉斯的好处是,写矢量作为功能f:V→R像@dkaeae,以及使用⟨⋅,⋅⟩对于通常内积,我们有由下式给出二次形式这非常好的表达L:
⟨f,Lf⟩=1d∑(u,v)∈E(f(u)−f(v))2.
A的最大特征值是d,对应于L的最小特征值,即0;第二最大特征值λ2的A对应的第二最小特征值L,这是1−λ2d。根据最小-最大原理,我们有
1−λ2d=min{⟨f,Lf⟩⟨f,f⟩:∑v∈Vf(v)=0,f≠0}.
注意⟨f,Lf⟩不,当我们交接班f由同一个常数每个顶点。因此,等效地,您可以为任何f:V→R定义“居中”函数f0乘f0(u)=f(u)−1n∑v∈Vf(v),和写
1−λ2d=min{⟨f,Lf⟩⟨f0,f0⟩:f not constant}.
⟨f0,f0⟩=1n∑{u,v}∈(V2)(f(u)−f(v))2n2
1−λ2d=min⎧⎩⎨⎪⎪2nd∑(u,v)∈E(f(u)−f(v))22n2∑{u,v}∈(V2)(f(u)−f(v))2:f not constant⎫⎭⎬⎪⎪.
What this means is that, if we place every vertex u of G on the real line at the point f(u), then the average distance between two independent random vertices in the graph (the denominator) is at most dd−λ2 times the average distance between the endpoints of a random edge in the graph (the numerator). So in this sense, a large spectral gap means that what happens across a random edge of G (local behavior) is a good predictor for what happens across a random uncorrelated pair of vertices (global behavior).