Maths Olympiad Prep

Track / Stage 6 / 158 of 400 #1158 of 1964

Problem 1158

National Olympiad, first round
Combinatorics Difficulty 6.2 Prove it

Lemma 6.5.2 (i) The mapping gg satisfies (F1) if and only if ff satisfies (F1);
(ii) The mapping gg is a circulation of G∗G^{*} if and only if ff satisfies (F1) and for every oriented graph C⃗\vec{C}, f(C⃗)=0f(\vec{C})=0.

This one wants a proof. Work it on paper, then read the official solution and mark yourself. Be honest about it: the record is only any use to you if it is.

Next problem →

Official solution

Prove that conclusion (i) follows from Lemma 6.5.1 (i) and the fact that e⃗↦e⃗∗\vec{e} \mapsto \vec{e} * is a bijection.
For the necessity part of (ii), we assume that gg is a flow on G∗G^{*} and consider a given orientation of a cycle C⊆GC \subseteq G. Let F:=E(C)F:=E(C). By Proposition 4.6.1, F∗F^{*} is a minimal cut in G∗G^{*}, i.e., for some appropriate X⊆V∗X \subseteq V^{*}, we have F∗=E∗(X,Xˉ)F^{*}=E^{*}(X, \bar{X}). According to the definitions of ff and gg, for one of the two orientations C⃗\vec{C} of CC, by Lemma 6.5.1 (ii) and Proposition 6.1.1, we get
f(C⃗)=∑e⃗∈C⃗f(e⃗)=∑d⃗∈E⃗∗(X,Xˉ)g(d⃗)=g(X,Xˉ)=0. f(\vec{C})=\sum_{\vec{e} \in \vec{C}} f(\vec{e})=\sum_{\vec{d} \in \vec{E}^{*}(X, \bar{X})} g(\vec{d})=g(X, \bar{X})=0.

Thus, according to f(C←)=−f(C⃗)f(\overleftarrow{C})=-f(\vec{C}), for the given orientation of CC, the corresponding value must also be zero.
For the sufficiency part, according to (i), we only need to prove that gg satisfies (F2). Here, we will prove a more general result: for every cut F∗=E∗(X,Xˉ)F^{*}=E^{*}(X, \bar{X}) of G∗G^{*}, we have g(X,Xˉ)=0g(X, \bar{X})=0. By Lemma 1.9.3, we can assume that F∗F^{*} is in fact a bond of G∗G^{*}. Since GG and G∗G^{*} are connected, FF is the edge set of a cycle C⊆GC \subseteq G (Proposition 4.6.1). According to Lemma 6.5.1 (ii), CC has an orientation C⃗\vec{C} such that {e⃗∗∣e⃗∈C⃗}=E∗→(X,Xˉ)\{\vec{e} * \mid \vec{e} \in \vec{C}\}=\overrightarrow{E^{*}}(X, \bar{X}). Therefore, according to the definitions of ff and gg, we have g(X,Xˉ)=f(C⃗)=0g(X, \bar{X})=f(\vec{C})=0.

With the help of Lemma 6.5.2, we can now prove the coloring-flow duality theorem for planar multigraphs. Let P=v0…vℓP=v_{0} \ldots v_{\ell} be a path with edges ei=vivi+1(i<ℓ)e_{i}=v_{i} v_{i+1} (i<\ell), then we denote (depending on the labeling of the vertices on PP)
P⃗:={(ei,vi,vi+1)∣i<ℓ}, \vec{P}:=\left\{\left(e_{i}, v_{i}, v_{i+1}\right) \mid i<\ell\right\},

and call P⃗\vec{P} a v0→vℓv_{0} \rightarrow v_{\ell} path (v0→vℓ\left(v_{0} \rightarrow v_{\ell}\right. path )). Similarly, PP is implicitly given by P⃗\vec{P}.

Source: NuminaMath-1.5, licensed Apache-2.0. Statement reproduced verbatim; metadata (topic, difficulty, ordering) added by this project.