下面发表在世界图论最权威杂志JCTb的泛圈图定理是18岁大学毕业、并在原北京大学分校现陕西理工大学2005年任副校长其后任正校长的张社民教授得到的(因做它的但不清楚就看一下这陕西理工大学见如陕西理工大学北京校友会会长张平院士、陕西理工大学北京校友会会长清华大学化学系教授李隽也多次候选院士、更有“总理李克强主持召开专家和企业家座谈会,苏剑、伍戈等专家和国家电力投资集团等企业负责人发了言”而第一个发言的苏剑也是陕西理工大学北京校友会会长,等等--看来这大学还是可以的。而这也算是18岁大学毕业的张社民校长做为陕西商州农村人在以前是不会有父母迫着读书的那这可说是天才还是因特殊爱好所致-可高考要考很多学科-谁能爱好那么多。当然早熟也非一切如华人唯一获得数学诺贝尔奖的丘成桐近18岁才读大学。如此张社民校长是我极佩服的数学专家之一,可在美国《数学评论》见天才的张校长2006年以前的几十年内也才仅有1篇论文-它就是下面我再给出简单证明的--这说明这世界三大难题之一的哈密顿图泛圈图非常不容易、也说明以前不容易。而且从我已竭尽全力的下面再证明中感到已极难创造出更好的技术理论方法等去给出这定理更好的证明
Short proof of well-known theorem on pancyclic graphs
Department of Mathematics,
Abstract:
In 1994, Zhang proved the
following well-known theorem: “Let G be a Hamiltonian graph of order n. If
there exists a vertex x such that d(x)+d(y)≥n for each y
not adjacent to x, then G is pancyclic or Kn/2,n/
In this paper,
we consider only finite, undirected graph without loops or multiple edges. For
a vertex u in G, we denote N(u)={v: uvÎE(G)} and G[N(u)] the subgraph induced by
N(u) in G.. Let Cm=x1x2…xmx1
denote a cycle of order m. Others notation can be found in [1].
In 1994, Zhang [3] proved the following
well-known theorem:
Theorem. Let G be a Hamiltonian graph of order n. If
there exists a vertex x such that d(x)+d(y)≥n for each y
not adjacent to x, then G is pancyclic or Kn/2,n/2.
Now we show a
simple proof of the above theorem.
Proof. Let Cn=x1x2…xnx1
be a Hamiltonian cycle of G,. We consider the following three cases.
Case
1. x1x3ÎE(G).
In this case, G
has C3. If G does not have Cj, j≥4, then x1xjÏE(G), and when
xi is adjacent to x1, then xi+1 is not
adjacent to xj for i+1≤j-1(if xi+1xjÎE(G), we have Cj=x1xi+1xi+2…xjxixi-1…x1,
a contradiction) or xi is not adjacent to xj for i≥j+1. Otherwise, we have Cj=x1x3x4…xjxix1.
Hence we can check that d(xj)≤n-d(x1)-|{xj}|,
this implies d(x1)+d(xj)≤n-1,
which contradicts the assumption of Theorem. The contradiction shows that G has
Cj, j=4, 5,…,i+1. Thus, G is pancyclic.
Case
2. x1x3ÏE(G) and x1x4ÎE(G).
We consider the
following. (i) When x1 is adjacent to some consecutive xi,xi+1.
If G does not have Cj, then x1xjÏE(G). (i-1)
When i+1≤j-1, xjx2ÏE(G)
(Otherwise, we have Cj=xjx2x3…xix1xi+1…xj).
Then for any xr, if x1xrÎE(G), then xr-1xjÏE(G) for r≤j-1, or xr-1xj,xr+1xjÏE(G) for r≥j+1( Otherwise, by x1x4ÎE(G), we easily
get Cj). Hence we can check that d(xj)≤n-d(x1)-|{x2}|, a contradiction.
(i-2) When j+1≤i,
then xjxi-1,xjxi,xjxi+1,xjxi+2ÏE(G). Hence d(x1)+d(xi)≤n-1, a contradiction. So we have Cj,j=5, 6,…n. And x1 is
adjacent to xi,xi+1 and x1x4ÎE(G), so we
also have C3, C4 , i.e., G is pancyclic.
(ii) When N(x1)∩{xh,xh+1…xi}={xh,xi},
i≥h+3. If G does not have Cj, then x1xjÏE(G). (ii-1)
When h+2≤j, then xjxh+1 or xjx2
or x1xj-1ÏE(G)( Otherwise, Cj=xjxh+1xh+2…xj-1x1xhxh-1…x2xj),
and by similar argument of (i), we have d(x1)+d(xj)≤n-1, a contradiction. (ii-2) When i-2≥j, then we have xjxi-1,xjxi+1ÏE(G). And by
similar argument of (i), we have d(x1)+d(xj)≤n-1, a contradiction. So G has Cj, j=5, 6,…n. Then when d(x1)≥n/2, x1 must be adjacent to consecutive xi,xi+1,
so G has C3; when d(x1)<n/2, by d(x1)+d(xj)≥n, d(xj)>n/2, so xj must be adjacent to consecutive xr,xr+1,
so G has C3, thus G is pancyclic.
(iii). G is the case without (i) and (ii). In this
case, clearly, d(x1)=n/2 and N(x1)={x2,x4,x6,x8,x2,x3,…x
Case
3. x1x3ÏE(G), x1x4ÏE(G).
Without loss of
generality, assume the most close vertex of x1 along Cn,
except x2 and xn, is xi(it is said that x3,x4…,xi-1
and xn-1,xn-2…,xn-(i-3) are not adjacent to x1).
For any j, if G does not have Cj, by similar argument of Case 1, xjxj-2ÏE(G). We also
have x1xj-1 or xjx3 or xjx2ÏE(G)(
otherwise, Cj=xjx3x4…xj-1x1x2xj).
And if xix1ÎE(G), then xi+1xj,ÏE(G) for i+1≤j-1, or xixjÏE(G) for j+1≤i. Hence we have d(x1)+d(xi)≤n-1, a contradiction. Thus, when j≥i, G has Cj, j=i,i+1,…n. When j≤i-1, by d(x1)+d(xj)≥n,
both x1,xj have a common neighborsÏ{x3,x4…,xi-1}, so G has Cj+1, j=3, 4,…,i-1. Again, since N(x1)∩{x2,x3,x4}={x2},
by similar argument of (ii) of Case 2, G has C3, so G is pancyclic.
References
1. J. A. Bondy, U. S. R. Murty,
Graph Theory with Applications, Macmillan, London and Elsevier,New York, 1976.
2. R. J.
Gould, Advances on the
Hamiltonian problem—a survey. Graphs Combin. 19 (2003),
1, 7--5
3. S. M.
Zhang, Pancyclism and bipancyclism of Hamiltonian
graphs. J. Combin. Theory Ser. B 60 (1994), 2,
159--168.
附注:Gould主席的著名综述文章“Advances
on the Hamiltonian problem—a survey. Graphs Combin. 19 (2003),
1, 7-
我们重视它是除了
为了探索尽可能多种最优方法,这里是我取得很大突破的若干著名方法备忘录:1、Whitney定理即“惠特尼定理” ;2、Bondy,定理;3、Bollobás定理;4、Faudree定理;5、Bermond定理;6、Linial定理;7、Hajnal定理;8、Kronk 定理;9、获得全国18项科技一等奖(数学唯一一项一等)的组合矩阵论中《中国科学》发表的柳-邵定理;10、范理事长的定理;11、邵理事长的定理
12、陈定理;13、海南的推广定理;14、宋理事长的定理;15、原北京大学分校张校长的定理,等等,当然并非简短的证明才值得学习,200页的证明虽头痛但很多方法可总结,我每次体会它们总有新的感想
下面发表在图论的世界权威杂志JCTb的二分图定理是18岁就已大学毕业、并是原北京大学分校现陕西理工大学正副校长、党委书记中唯一有博士学位的张社民校长得到的,但他的证明非常复杂,如此我下面6行字就证明(18岁大学毕业这在以前是天才,如陈后华人第一数学大师丘成桐近18岁才读大学。可在美国《数学评论》见天才的张校长2006年以前也仅有1篇论文,可见哈密顿图之艰难--如他这唯一论文的被我下面六行字证明的结果就不易-有些人要经百次、千次、有些万次失败--可能没想到这六行字会有上万次选择的可能吧?而许多接近正确的错误排除一次都不容易
1. Gould主席的著名综述文章“Advances on the Hamiltonian problem—a survey. Graphs
Combin. 19
(2003), 1, 7-
2. 张社民教授对上面定理的证明共引用3个引理,其中之一是上面较复杂的引理4.2(Lemma 4.2)(较复杂也如张社民教授在论文中陈述这一引理就用9行字,而我上面对定理的证明却仅用6行字)。
也就是我上面仅用六行字就证明
可参考二分图专著:Bipartite
Graphs and their Applications,此书3个作者Armen Asratian于1980年获得莫斯科大学博士,Tristan. Denley现任美国Austin
Peay州立大学常务校长,Roland Häggkvist于1977年获得Dirac的博士。其实象我们大学毕业于图论的研究生林越等老师都能知道六行字就证完的关键在哪?
为了探索尽可能多种最优方法,这里是我取得很大突破的若干著名方法备忘录:获得国家奖的范定理、Whitney定理、1991年《中国科学》发表的柳-邵定理、陈定理、L猜想、宋理事长的宋定理、Bondy,定理、Bermond定理、Linial定理、海南省最重要的黄定理,等等,当然并非简短的证明才值得学习,200页的证明虽头痛但很多方法可总结,我每次体会它们总有新的感想
1. Gould主席的著名综述文章“Advances on the Hamiltonian problem—a survey. Graphs
Combin. 19
(2003), 1, 7-
2. 上面张社民教授证明的上面定理共引用3个引理,其中之一是上面较复杂的引理4.2(Lemma 4.2)(较复杂是这引理在张社民教授的论文中也是九行字,而我上面对定理的证明仅六行字)它是加州最悠久的圣何塞州立大学Edward Schmeichel教授和这里第2段说80年代是系主任的John Mitchem合作的论文Bipartite
graphs with cycles of all even lengths. J. Graph Theory 6 (1982), 4, 429-439中的非常复杂的结果。
我上面仅用六行字证明
为了探索多种最优方法,这里是我取得突破很大的若干著名方法备忘录::1、Whitney定理即“惠特尼定理” ;2、Bondy,定理;3、Bollobás定理;4、Faudree定理;5、Bermond定理;6、Linial定理;7、Hajnal定理;8、Kronk 定理;9、获得全国18项科技一等奖(数学唯一一项一等)的组合矩阵论中《中国科学》发表的柳-邵定理;10、范理事长的定理;11、邵理事长的定理
12、陈定理;13、海南的推广定理;14、宋理事长的定理;15、原北京大学分校张校长的定理,等等,当然并非简短的证明才值得学习,200页的证明虽头痛但很多方法可总结,我每次体会它们总有新的感想